framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,2560,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,5120,0.038839111725489296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,3072,0.015047111445003085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,10240,0.014120888378885059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,512,0.005579555614127054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,512,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,2048,0.007618666523032718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,512,0.005050666630268097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,128,0.0018622221218215094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,3072,0.27271556854248047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,256,0.002834666727317704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,3072,0.011186666786670685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,1024,0.004940444396601783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,3584,0.007383999725182851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,6144,0.01311377767059538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,768,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,2560,0.011944000091817645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,10240,0.05046222276157803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,2048,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,1536,0.006344888773229387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,8192,0.009538667069541084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,3584,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,6144,0.02366133365366194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,2560,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,10240,0.03334488802485996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,256,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,2048,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,65536,3.2067627376980252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,5120,0.011343999869293638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,16384,0.4049679968092177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,7168,0.00499911109606425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,3584,0.010656889114114972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,2048,0.042437334855397545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,768,0.005236444373925527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,16384,0.059270222981770836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,5120,0.018325333793958027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,128,0.006468444648716185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,1024,0.011743110915025076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,16384,0.013915555344687568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,10240,0.043078223864237465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,2560,0.01593866613176134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,2048,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,3072,0.007677333222495184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,64,0.008291555775536431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,1024,0.010450666977299584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,1024,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,8192,0.8411155806647407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,12288,0.018231999542978074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,256,0.0069146665434042616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,128,0.015808888607554965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,128,0.0021742222209771476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,1024,0.04761244522200691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,512,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,10240,0.017436444759368896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,64,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,2048,0.006984000404675801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,4096,0.008289777570300633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,6144,0.07299288776185778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,16384,0.023416888382699754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,7168,0.009992889232105678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,256,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,4096,0.017813333206706576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,4096,0.008681777450773451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,768,0.0029226665695508323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,2048,0.009200000100665623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,10240,0.017495110630989075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,12288,0.020463110672103036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,2560,0.03308711118168301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,10240,0.05404177639219496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,128,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,64,0.007335999773608313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,1024,0.012442666623327466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,1536,0.00665155549844106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,3584,0.01606666710641649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,1536,0.1935715542899238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,3072,0.010074666804737514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,16384,0.026543110609054565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,128,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,2048,0.08917511171764797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,2048,0.022295110755496558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,512,0.017695110705163743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,512,0.007998222278224098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,256,0.02228533393806881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,16384,0.02442933287885454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,512,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,256,0.002491555487116178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,64,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,3072,0.012796444197495779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,65536,0.058238221539391406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,256,0.002471111093958219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,65536,0.06887644529342651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,64,0.00888888869020674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,8192,0.013838222457302941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,1536,0.029478222131729126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,32,0.02676444417900509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,4096,0.00452622233165635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,7168,0.03988533218701681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,512,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,2560,0.017892445127169292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,512,0.0035208890007601846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,2560,0.03230844603644477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,2560,0.009511999785900116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,1536,0.016209777858522203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,1024,0.014783110883500842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,1536,0.0042071110672420925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,5120,0.015822221835454304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,256,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,128,0.009642666412724389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,5120,0.061315556367238365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,3072,0.06494666470421685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,2560,0.08274755875269572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,2048,0.012653333445390066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,64,0.0021377777059872947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,16384,0.08876622385448879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,3584,0.008048888709810045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,5120,0.004868444469239977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,5120,0.021726222501860723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,10240,0.018210666047202218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,3072,0.0269128895468182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,5120,0.053802665736940175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,5120,0.02072533302836948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,512,0.00833511104186376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,1024,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,2560,0.016843555702103507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,12288,0.03195022212134467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,3584,0.018912000788582694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,16384,0.015816888875431485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,256,0.014780445231331719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,1536,0.015471999843915304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,8192,0.0074684446056683855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,256,0.0021804444905784395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,5120,0.011050666371981302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,1024,0.013752889302041797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,3072,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,65536,0.06087733639611138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,32,0.002587555597225825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,7168,0.01204177737236023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,64,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,1536,0.027443556321991816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,128,0.012468444804350534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,8192,0.0739884442753262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,256,0.002475555572244856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,2048,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,12288,0.05692444245020548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,1024,0.033117334047953285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,512,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,5120,0.01592266725169288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,10240,0.05431911018159655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,768,0.01482222146458096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,1536,0.019711999429596793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,16384,0.016136889656384785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,128,0.014563555518786112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,2560,0.0068915556702348925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,768,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,3072,0.06232177548938327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,3072,0.03356000118785434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,65536,0.09413688712649876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,1024,0.02195022172398037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,32,0.002187555655837059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,16384,0.015815999772813585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,3584,0.009301332963837517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,5120,0.009708444277445475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,6144,0.15312088860405818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,8192,0.2588035530514187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,32,0.0018524444765514797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,5120,0.016903999778959487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,64,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,7168,0.03123289015558031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,32,0.014446222119861178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,6144,0.013491555220550962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,128,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,4096,0.038154668278164335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,1536,0.017510222064124215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,12288,0.02313066687848833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,3584,0.07984977960586548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,3072,0.015812445018026564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,10240,0.03390044305059645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,8192,0.02240444388654497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,2048,0.011393778026103973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,65536,0.16349155373043484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,32,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,2560,0.019305777218606737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,10240,0.08765955766042073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,16384,0.027492443720499676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,10240,0.034232887956831194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,7168,0.024878222081396315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,128,0.005411555369695027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,2560,0.02609866691960229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,5120,0.0424124432934655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,1024,0.016576889488432143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,768,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,3584,0.022720000810093347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,128,0.005205333232879639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,2048,0.00740622232357661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,4096,0.007697777615653143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,2048,0.0162471118900511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,3584,0.009684444301658207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,12288,0.06373155779308744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,5120,0.015321777926550971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,2048,0.0247688889503479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,65536,0.06185866726769341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,2560,0.01236088905069563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,6144,0.05250666538874308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,4096,0.011287111375066968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,256,0.011019555230935415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,12288,0.013450666434235044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,2560,0.015196444259749519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,32,0.0018399999373488957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,512,0.0031644445326593188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,10240,0.10665422015719944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,7168,0.009761778016885122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,4096,0.017862222260899015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,7168,0.016738666428460013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,65536,0.07856622007158068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,512,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,65536,0.06592355834113227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,256,0.011361777782440186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,256,0.009376889301670922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,2560,0.026399110754330952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,65536,0.058092441823747426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,16384,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,7168,0.01015911085738076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,64,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,32,0.001999999913904402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,128,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,512,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,8192,0.041088889042536415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,3072,0.01940622263484531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,10240,0.012381333443853589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,512,0.01651733285850949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,768,0.008388444781303406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,512,0.013073777986897362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,7168,0.018046221799320646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,2048,0.006882666713661618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,7168,0.014418666561444601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,65536,0.07860444651709662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,4096,0.00832088871134652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,256,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,256,0.006984000404675801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,10240,0.017872888180944655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,1024,0.018937778141763475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,2048,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,5120,0.01588088936275906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,3072,0.015909334023793537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,1536,0.004315555509593752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,10240,0.007043555378913879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,3584,0.016579555140601266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,2048,0.006337777607970768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,2048,0.00666222224632899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,64,0.01555999947918786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,12288,0.03313066562016805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,3584,0.024886222349272832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,256,0.00600000015563435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,16384,0.07794488800896539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,8192,0.014051554931534661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,16384,0.09449778000513713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,10240,0.02297244469324748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,128,0.0021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,10240,0.01647733317481147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,16384,0.04832266767819723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,768,0.002507555608948072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,512,0.11370133029090033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,64,0.0028719999310043124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,768,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,512,0.015136000182893542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,16384,0.07610399855507745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,2560,0.009670221971140968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,2560,0.004569777597983678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,4096,0.01977066695690155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,7168,0.009900444083743626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,6144,0.01479111115137736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,6144,0.029927998781204224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,2560,0.011755555868148804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,768,0.017312889297803242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,3072,0.022590221630202398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,2048,0.02424977719783783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,4096,0.007790221936172909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,256,0.03581600056754218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,512,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,768,0.008558221989207799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,4096,0.007994666695594788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,16384,0.023027555810080633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,6144,0.017151999804708693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,1024,0.002943111169669363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,1536,0.06308622492684259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,10240,0.17445777522193062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,2560,0.014152000347773233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,3072,0.017553778158293832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,2048,0.010701333483060202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,2048,0.006512889017661412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,2560,0.042133334610197276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,512,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,8192,0.02113155523935954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,65536,0.13845511277516684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,2560,0.013738666971524557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,3584,0.013467555244763693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,5120,0.008417777717113495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,2048,0.028791109720865887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,2048,0.006647111227114995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,8192,0.1733244392606947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,768,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,512,0.4079768922593858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,8192,0.012427555190192329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,3072,0.1320213344362047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,512,0.005655111124118169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,3072,0.004914666629499859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,256,0.017887110511461895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,512,0.013439999686347114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,32,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,16384,0.019845333364274766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,12288,0.027760000692473516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,4096,0.011076444553004371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,1024,0.012415111064910889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,2560,0.007423111134105259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,12288,0.029550221231248643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,512,0.0028719999310043124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,3072,0.08950755331251357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,6144,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,3072,0.004247111164861255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,512,0.017166222135225933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,5120,0.21427378389570448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,1024,0.0069102222720781965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,64,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,16384,0.02075199948416816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,6144,0.0357893341117435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,3584,0.013384000294738345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,1536,0.01721333298418257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,1536,0.004581333448489507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,3072,0.08579822381337483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,65536,0.13243022229936388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,16384,0.049327999353408813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,65536,0.15210133128696018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,1536,0.006308444258239534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,6144,0.05444444550408257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,2560,0.01960711181163788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,1024,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,3584,0.014452444182501899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,4096,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,8192,0.011715555356608497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,10240,0.015823110938072205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,65536,0.050273779365751475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,12288,0.07839466465844049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,128,0.015134221977657743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,32,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,768,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,16384,0.045309331681993276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,65536,0.12470666567484538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,12288,0.034974220726225115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,512,0.01958755486541324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,1024,0.017779555585649278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,2048,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,8192,0.029215110672844782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,256,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,10240,0.027592000034120347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,1536,0.016215110818545025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,3584,0.0925111108356052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,8192,0.045370668172836304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,2048,0.06546844376458062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,7168,0.011356444822417365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,512,0.01718755563100179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,10240,0.1287955575519138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,1536,0.6022675302293565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,6144,0.01822933389080895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,65536,0.8712959819369845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,65536,0.16731377442677817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,1024,0.003548444559176763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,512,0.01667022208372752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,2048,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,256,0.010029333333174387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,32,0.005534222142563925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,5120,0.015972443752818637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,7168,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,1536,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,12288,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,3072,0.007662222617202335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,2560,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,65536,0.08148622512817383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,3072,0.03753688931465149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,12288,0.06407199965582953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,64,0.006173333360089197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,512,0.029669334491093952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,512,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,3072,0.009335999687512716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,3072,0.009349333743254343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,1024,0.015460444821251763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,16384,0.030055999755859375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,2560,0.0653360022438897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,128,0.013771555489963956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,8192,0.07086133294635348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,128,0.0018648888087934917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,1024,0.01479555500878228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,10240,0.014323555760913424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,65536,0.2802062299516466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,128,0.0324888891643948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,4096,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,3584,0.007341333561473423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,128,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,10240,0.037534223662482366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,2048,0.02029244436158074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,256,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,256,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,256,0.0021404443929592767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,10240,0.03889688849449158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,6144,0.029256890217463177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,2048,0.025570667452282373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,512,0.06715199682447645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,3072,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,64,0.00184888889392217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,5120,0.020815999971495736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,16384,0.0174319992462794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,8192,0.025748444928063288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,768,0.005947555518812603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,64,0.0028231110837724474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,10240,0.02737155556678772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,3072,0.012472000386979846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,8192,0.20366577307383218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,5120,0.012857777376969656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,65536,0.06453244553671943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,10240,0.012944888737466602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,512,0.0025111111915773815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,1536,0.024477332830429077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,4096,0.01573333309756385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,1024,0.01717511150572035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,768,0.005715555614895291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,8192,0.06259910927878486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,1024,0.01822933389080895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,1024,0.020106666617923312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,1024,0.008012444608741337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,1024,0.005629333357016246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,768,0.017497777938842773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,6144,0.10589422119988336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,64,0.0016293333222468693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,5120,0.05031999945640564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,65536,0.11188088523017035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,10240,0.011703111231327057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,1536,0.010119111173682744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,768,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,3584,0.013144888811641268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,8192,0.0340302222304874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,1024,0.15006044175889757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,3072,0.016139555308553908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,5120,0.09138311280144586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,65536,0.06480444139904447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,2560,0.015444444285498725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,1536,0.009879111415810054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,16384,0.09085510836707221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,10240,0.013059555656380124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,7168,0.013372444444232516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,8192,0.01107377807299296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,512,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,3584,0.017204445269372728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,512,0.017683555682500202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,128,0.0031635556370019913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,16384,0.0163111107216941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,2560,0.004570666700601578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,8192,0.008620444271299574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,10240,0.018576888574494254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,768,0.005584888988071018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,16384,0.032594667540656194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,6144,0.02200977835390303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,1024,0.005973333285914526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,10240,0.03895733422703213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,2560,0.008700444466537898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,65536,0.020081778367360432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,256,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,4096,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,10240,0.03524533245298598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,10240,0.009698666632175446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,256,0.009944000177913243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,65536,0.06579911046557956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,1024,0.0069288888739215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,5120,0.1162746614880032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,1536,0.008477333519193862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,512,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,1024,0.01648888819747501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,768,0.0024844445288181305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,128,0.0018604445374674266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,5120,0.01258133351802826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,256,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,5120,0.006276444428496891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,12288,0.03738666574160258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,65536,0.07759110795127021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,1024,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,12288,0.03759200043148465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,3072,0.011048888994587792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,65536,0.15508711338043213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,8192,0.014368888404634265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,256,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,2560,0.010266666611035665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,3072,0.010337777435779572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,64,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,65536,0.0599493318133884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,65536,0.03316888875431485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,1024,0.013072888884279462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,7168,0.03944622145758735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,4096,0.0105262224872907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,12288,0.004945777770545748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,256,0.0042133331298828125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,256,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,3072,0.030943112240897283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,8192,0.030015110969543457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,3072,0.008779555559158325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,768,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,65536,0.1886204481124878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,768,0.00282222218811512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,1536,0.004594666676388847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,10240,0.08762311273150974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,256,0.01516533394654592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,65536,0.09174399905734593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,10240,0.012758221891191272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,512,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,2048,0.029142220815022785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,7168,0.030877331892649334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,10240,0.04167822334501478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,65536,0.07643910911348131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,2048,0.008047999607192146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,3584,0.026495110657480028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,64,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,8192,0.017841777867741056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,5120,0.01273155543539259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,32,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,768,0.0028115556471877625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,2560,0.033906665113237165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,3584,0.010747555229398938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,65536,0.030768000417285498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,6144,0.10088800059424506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,7168,0.03621422251065572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,7168,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,65536,0.09394399987326728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,128,0.006561777657932705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,4096,0.008191111187140146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,2048,0.0069155556460221606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,128,0.005879999862776862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,4096,0.04127911064359877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,2048,0.02386577758524153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,3072,0.010328888893127441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,65536,3.254392835828993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,3584,0.019941333267423842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,32,0.0018133332745896445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,12288,0.027257778578334387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,32,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,7168,0.35449332661098903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,16384,0.02402399977048238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,1024,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,512,0.013658666776286231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,3584,0.025715554753939312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,512,0.006561777657932705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,16384,0.05285155442025927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,5120,0.023494222097926672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,3072,0.03702311052216424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,6144,0.017906667457686532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,2048,0.013422222601042854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,2560,0.030849777989917334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,12288,0.013407111167907715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,6144,0.01714933415253957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,3072,0.03024622135692173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,7168,0.3995030985938178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,8192,0.043079998758104115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,6144,0.013408000270525614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,512,0.006593777901596493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,1024,0.0028275555620590844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,12288,0.011408888631396823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,7168,0.029784000582165186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,5120,0.05202755663130018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,7168,0.033160001039505005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,12288,0.04064711266093784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,8192,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,3584,0.00795911086930169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,4096,0.008625778059164682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,5120,0.020974222156736586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,5120,0.006650666809744305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,7168,0.03545511100027297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,4096,0.029088000456492107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,768,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,2560,0.030431111653645832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,4096,0.024992000725534227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,4096,0.0092613332801395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,6144,0.07097599903742473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,8192,0.1348302231894599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,3072,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,7168,0.11354577541351318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,2560,0.016672000288963318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,3584,0.013255111045307584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,256,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,4096,0.007808888951937358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,12288,0.06579999791251288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,256,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,6144,0.019573332534896005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,12288,0.07500800159242418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,2048,0.008006222546100616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,256,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,10240,0.01237066669596566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,7168,0.01752088963985443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,2560,0.018235555953449674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,64,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,65536,0.058428446451822914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,512,0.015374221735530429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,64,0.003972444269392225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,65536,0.06292266978157891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,65536,0.14698933230506048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,7168,0.031082666582531396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,2048,0.010468444062603844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,1024,0.012763555679056378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,65536,0.0703164471520318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,2048,0.09100888835059272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,1024,0.03705599904060364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,65536,0.17023821671803793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,768,0.01687111126052009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,128,0.010029333333174387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,5120,0.0236791107389662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,5120,0.030200001266267564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,2560,0.006955555743641323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,512,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,512,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,768,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,2560,0.004880889008442561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,1536,0.01239466667175293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,3072,0.062000890572865806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,512,0.0035173334181308746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,8192,0.012682666381200155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,16384,1.279354625278049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,64,0.039966222312715315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,512,0.018458666072951425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,5120,0.016043555405404832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,4096,0.018964444597562153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,12288,0.029106666644414265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,65536,0.21778310669793022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,128,0.002163555473089218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,5120,0.008712000317043727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,12288,0.02780088782310486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,7168,0.0629955530166626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,256,0.005991111199061076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,3072,0.024337778488794964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,5120,0.009420444567998251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,7168,0.006699555450015598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,4096,0.007962666451931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,3584,0.007967111137178209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,4096,0.013448889056841532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,65536,0.1000559992260403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,1536,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,10240,0.026734222968419392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,1024,0.003871110992299186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,7168,0.014380445082982382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,2560,0.014642667439248828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,6144,0.013771555489963956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,1536,0.030032889710532293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,768,0.003931555483076307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,64,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,10240,0.1629991134007772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,256,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,16384,0.021000888612535264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,2048,0.08248533142937554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,8192,0.2836693392859565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,5120,0.27454132503933376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,512,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,7168,0.02055999967787001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,1024,0.003148444410827425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,6144,0.016937777400016785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,5120,0.02479733361138238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,1024,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,64,0.005631111148330901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,3584,0.01404711107412974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,768,0.029861334297392104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,10240,0.008757332960764566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,3072,0.010105777945783403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,3584,0.02218933403491974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,2560,0.011564444336626263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,65536,0.6985893249511719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,12288,0.01718133356836107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,5120,0.009004444711738164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,128,0.00249777775671747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,32,0.006253333141406377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,6144,0.0074471111098925276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,64,0.0018257778137922287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,1024,0.003519999898142285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,1024,0.0069573331210348345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,3072,0.03220266766018338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,7168,0.060511112213134766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,10240,0.03052622079849243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,16384,0.04808888832728068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,4096,0.017290666699409485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,1536,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,768,0.005965333431959152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,768,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,768,0.01716711123784383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,12288,0.02830222249031067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,6144,0.012656889028019376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,65536,0.06984711355633207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,1024,0.0031679999083280563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,8192,0.014422222971916199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,7168,0.014485332700941296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,256,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,3584,0.020267556111017864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,7168,0.06120710902743869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,65536,0.8261422051323785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,8192,0.025589333640204534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,2048,0.009995555712117089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,64,0.0026480000880029467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,3072,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,8192,0.03070488903257582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,256,0.005325333525737126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,8192,0.04835822184880575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,3072,0.007249777515729268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,6144,0.027105778455734253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,3072,0.05717244413163927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,6144,0.036637332704332136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,2048,0.012726222475369772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,512,0.007293333609898885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,768,0.016157334049542744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,768,0.0033760000434186724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,128,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,2048,0.011603555745548673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,1024,0.015599111715952555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,6144,0.009328888522254096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,256,0.009335999687512716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,2048,0.010994666980372535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,3584,0.02678844498263465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,5120,0.012720000412729053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,3072,0.00795733349190818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,128,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,64,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,3072,0.03182933396763272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,4096,0.014123555686738757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,256,0.013807111316257052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,6144,0.013120888835854001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,3584,0.020089776979552377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,2560,0.010378667049937779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,768,0.03563377923435635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,2560,0.020801777640978496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,128,0.005621333503060871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,1536,0.009293333523803288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,6144,0.006776000062624614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,512,0.0029253332565228143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,1536,0.00833333366447025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,256,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,512,0.006907555378145642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,12288,0.027969777584075928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,1024,0.013810666898886362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,512,0.006792888873153263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,12288,0.020244444410006206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,4096,0.0092604441775216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,10240,0.01236177815331353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,5120,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,256,0.0028284444577164124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,16384,0.020936000678274367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,1536,0.008871999879678091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,1024,0.00318755561278926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,1024,0.014853333433469137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,8192,0.16965066062079537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,2560,0.007033777733643849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,8192,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,2560,0.017304889029926725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,2048,0.06649333238601685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,64,0.008279110822412703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,10240,0.007920000288221572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,32,0.026870222555266485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,256,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,3584,0.016407110624843173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,8192,0.1379964484108819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,512,0.01348355578051673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,5120,0.046801779005262584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,7168,0.08114666408962674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,8192,0.010743111371994019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,12288,0.12705066468980578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,512,0.014097778333557976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,65536,0.024480000138282776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,12288,0.017521777086787753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,65536,5.854424794514974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,1536,0.01037333326207267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,6144,0.010202666951550378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,32,0.0038524443904558816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,6144,0.012130666938092975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,512,0.007254222200976477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,4096,0.023476445012622412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,128,0.005031110925806893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,6144,0.007977777885066139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,64,0.0018231111268202465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,768,0.011216888825098673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,7168,0.035755554835001625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,512,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,2560,0.03315022256639268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,256,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,128,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,12288,0.029105779197480943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,5120,0.009382222261693742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,8192,0.009682666924264696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,2048,0.005506666584147348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,4096,0.015287111202875773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,64,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,16384,0.020948444803555805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,7168,0.03458844290839301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,2048,0.006974221931563483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,32,0.0053182223604785065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,10240,0.01719111038578881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,3584,0.016876444220542908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,32,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,3584,0.013564444250530668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,768,0.00693866651919153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,3072,0.007635555333561367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,32,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,7168,0.01108000013563368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,1536,0.019290667441156175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,768,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,5120,0.5289884673224555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,128,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,1536,0.011376000112957425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,1536,0.006953777538405524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,512,0.012123555772834353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,5120,0.013166222307417126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,2560,0.011122666299343109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,7168,0.019991111424234178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,64,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,2560,0.020803555846214294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,4096,0.018426666657129925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,1536,0.005963555640644497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,128,0.002492444382773505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,4096,0.008000888758235509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,2048,0.0097884444726838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,768,0.007027555671003129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,256,0.012178666889667511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,2560,0.008736888567606607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,5120,0.016546666622161865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,65536,0.10129777590433757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,65536,0.06399555338753594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,512,0.009000889129108852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,5120,0.009000889129108852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,12288,0.02640266716480255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,2560,0.008817777865462834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,2560,0.05993600024117363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,3072,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,7168,0.00998755544424057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,12288,0.029495110114415485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,3072,0.11469066143035889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,8192,0.3562293317582872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,128,0.01794577803876665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,1024,0.006010666489601135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,128,0.010323555933104621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,65536,0.39000709851582843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,32,0.005625777774386936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,7168,0.01348000019788742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,1024,0.0028977776981062363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,8192,0.3072311083475749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,8192,0.015360888507631091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,2560,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,64,0.008643555144468943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,12288,0.03981066743532816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,2560,0.029419554604424372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,128,0.0018151110659042995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,32,0.0027964444210131965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,10240,0.022628444764349196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,3072,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,3072,0.008634666601816813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,16384,0.7804320123460559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,256,0.03458844290839301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,2560,0.03341866532961527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,7168,0.010319111247857412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,10240,0.014160000615649752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,65536,0.12051555845472549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,5120,0.011712888876597086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,1536,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,2560,0.01921688848071628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,3072,0.008347555167145198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,3584,0.014774221513006421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,10240,0.044489777750439115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,16384,0.023099554909600153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,8192,0.0247697780529658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,768,0.01238222254647149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,7168,0.017712000343534682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,12288,0.0195448895295461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,7168,0.03623822331428528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,16384,0.016145777371194627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,3584,0.014825777875052558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,7168,0.07734400033950806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,768,0.0031759999692440033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,12288,0.21949511104159883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,4096,0.02558044426971012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,8192,0.034266667233573064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,65536,0.07580177651511298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,8192,0.030065778228971694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,5120,0.02130844526820713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,768,0.007584888901975419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,512,0.013901333014170328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,128,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,2048,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,512,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,128,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,12288,0.018107554978794523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,256,0.01682400041156345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,768,0.01680888897842831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,1536,0.015456888410780164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,10240,0.016163556112183463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,8192,0.12771111064487034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,6144,0.004918222212129169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,3072,0.023375110493765935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,7168,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,3072,0.02936444348759121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,128,0.00565244464410676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,65536,0.08856266736984253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,768,0.029818667305840388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,16384,0.07896622021993001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,3584,0.007344889144102733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,768,0.0024871110088295406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,8192,6.270717196994358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,512,0.0025235555238193935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,2560,0.004564444637960858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,3584,0.012016889121797351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,1024,0.051854223012924194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,3584,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,1024,0.01515111161602868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,5120,0.048934221267700195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,768,0.0761715571085612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,256,0.01383733335468504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,2048,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,3584,0.009959999885823991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,10240,0.012073777616024017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,2560,0.011343999869293638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,768,0.007863110966152614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,3072,0.021241777473025854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,65536,0.06322044134140015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,256,0.014550222290886773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,512,0.01310222182008955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,256,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,16384,0.02077244387732612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,1536,0.0032666667054096856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,16384,0.021985777550273474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,5120,0.004874666531880696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,4096,0.007982222570313348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,1536,0.019495111372735765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,10240,0.043771555026372276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,1536,0.027098667290475633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,5120,0.008681777450773451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,3584,0.01906755566596985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,10240,0.01240711079703437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,256,0.014108444253603617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,1536,0.006313777632183499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,7168,0.029976000388463337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,10240,0.012051555845472546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,16384,0.07579911417431302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,512,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,512,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,2048,0.012055111428101858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,16384,0.02829244401719835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,1536,0.005956444475385878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,10240,0.016384000579516094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,768,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,128,0.015626667274369132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,256,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,1536,0.00481599983241823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,7168,0.03238666719860501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,4096,0.011192888849311404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,6144,0.01016444464524587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,768,0.006912888752089606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,256,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,32,0.0018364443547195857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,512,0.0025191110455327565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,256,0.012467555701732635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,16384,0.054639998409483165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,1024,0.003168888803985384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,1024,0.006225777582989798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,3584,0.021171554923057556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,256,0.07161777549319796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,2560,0.010089777410030365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,64,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,12288,0.011571555501884885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,65536,0.13671199480692545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,128,0.016363556186358135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,3072,0.005335999859703912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,512,0.012378666963842181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,65536,0.15155022674136692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,2560,0.004920000003443824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,16384,0.06955555412504408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,2048,0.008983111215962304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,7168,0.06978044244978163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,6144,0.018212444252438016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,12288,0.05028088887532552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,128,0.008623110751310984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,2048,0.009343110852771336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,6144,0.004951999833186467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,256,0.005894222193294101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,2048,0.013813333378897773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,768,0.017172444197866652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,1024,0.019880000087949965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,12288,0.09565955400466919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,16384,0.16687111059824625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,5120,0.00488800017370118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,128,0.0052800000541739995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,6144,0.014919110470347934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,10240,0.21947288513183594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,768,0.01794755624400245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,256,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,1536,0.009687110781669617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,5120,0.004886222382386525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,3584,0.019156444403860304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,768,0.01688799924320645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,3584,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,3584,0.006774222271309958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,768,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,2560,0.0052284445199701525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,1024,0.006621333460013072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,7168,0.027632888820436265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,2048,0.006583111153708563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,128,0.0028497777465316984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,3072,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,3584,0.007237333390447829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,64,0.003141333245568805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,64,0.0021386666016446217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,10240,0.0846995578871833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,256,0.015513777732849121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,6144,0.03189422355757819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,7168,0.027477333943049114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,3584,0.06846488846672906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,256,0.014489778214030795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,16384,0.012849777936935425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,32,0.001763555531700452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,2048,0.01370933320787218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,512,0.002469333302643564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,16384,0.22315645217895508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,5120,0.010599111517270407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,16384,0.01665599975321028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,12288,0.01796533332930671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,2560,0.008962666822804345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,12288,0.06432977649900648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,3584,0.013380444712109037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,768,0.01995111174053616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,2560,0.01700177788734436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,768,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,7168,0.04484622346030342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,16384,0.021371554997232225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,32,0.006144000010357962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,3584,0.00795999997191959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,65536,0.3071671062045627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,10240,0.2292257679833306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,256,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,128,0.011765333513418833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,2048,0.023246222072177466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,256,0.006221333311663733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,6144,0.009974222216341231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,65536,0.07265688975652058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,32,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,1024,0.028775112496482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,65536,0.6631253560384115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,32,0.11989955107371013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,16384,0.012081777883900536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,3072,0.009311999711725447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,1536,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,768,0.009658666948477427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,1024,0.05021599928538004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,128,0.017165333032608032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,12288,0.05969422393374973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,512,0.0144177774588267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,6144,0.023000889354281958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,1024,0.016175111134847004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,1024,0.025081778566042583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,1024,0.003140444349911478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,5120,0.012784000072214337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,4096,0.007992888490358988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,1024,0.002834666727317704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,6144,0.019581332802772522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,256,0.0053288886944452924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,65536,0.1279946698082818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,768,0.003850666599141227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,6144,0.014855111638704935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,1536,0.012087999946541257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,2048,0.04775999983151754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,512,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,16384,0.05113510953055488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,1024,0.0031724443866146933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,16384,0.13990044593811035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,10240,0.04174044397142199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,128,0.014775110615624322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,64,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,65536,0.07950844367345174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,32,0.006406222366624408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,2560,0.01755466726091173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,2560,0.018547554810841877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,4096,0.008017777568764156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,7168,0.010069333016872406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,4096,0.007630222373538547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,3072,0.012506666282812754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,2048,0.012730666332774691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,512,0.006545777950021956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,768,0.015511110424995422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,4096,0.010728889041476779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,65536,0.15457777182261148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,5120,0.008709333009190029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,1024,0.015150222513410779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,65536,0.13680977291531032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,1536,0.008177777959240807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,64,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,5120,0.00832444429397583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,10240,0.017655111021465726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,3072,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,1024,0.007749333149856991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,4096,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,3072,0.01828444500764211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,3584,0.009900444083743626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,512,0.015557333827018738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,5120,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,5120,1.9078621334499781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,7168,0.02472355630662706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,65536,0.07419200075997247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,1024,0.004882666799757215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,32,0.006631999793979857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,5120,0.09133066733678182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,12288,0.025989333788553875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,16384,0.3714275625016954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,1536,0.010075555907355415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,3072,0.015497777197096082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,1024,0.016544888416926067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,4096,0.011018666956159802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,12288,0.04235733217663235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,2048,0.014575999644067554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,512,0.014471999473041959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,7168,0.008937777744399177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,12288,0.007689778175618913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,6144,0.01313333296113544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,12288,0.01776622235774994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,1024,0.007728888756699032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,2560,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,1024,0.006235555642180973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,65536,0.059009777175055615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,64,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,12288,0.07489689191182454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,2560,0.053038222922219165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,2048,0.006795555353164673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,512,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,5120,0.1298888921737671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,16384,0.01628266606065962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,12288,0.04960088928540548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,12288,0.013785777820481194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,128,0.001871111078394784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,64,0.005960888746711943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,4096,0.05256533291604784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,6144,0.03394133183691237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,6144,0.014204444156752693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,64,0.0019288888821999233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,8192,0.021393777595625982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,1024,0.00665866666369968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,16384,0.034224000242021345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,64,0.0034959999223550162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,65536,0.12139288584391277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,128,0.005212444398138259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,8192,0.01571733256181081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,768,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,7168,0.013047110703256396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,1024,0.006253333141406377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,6144,0.020238222347365487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,256,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,1536,0.03596266773011949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,768,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,6144,0.009696000152164036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,5120,0.017290666699409485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,4096,0.010040000081062317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,2048,0.05215200119548374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,3072,0.017490666773584154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,65536,0.06565333074993558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,3584,0.02147644427087572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,3584,0.007673777639865875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,4096,0.015089778436554803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,7168,0.012739555703269111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,2048,0.26494399706522626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,12288,0.3948426776462131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,1024,0.006510222123728857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,16384,0.03408177693684896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,12288,0.01789866718981001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,2560,0.20240267117818198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,12288,0.024047111471494038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,16384,0.05353600117895338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,8192,0.014488000008794995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,65536,4.431868235270183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,512,0.024247111545668706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,32,0.008558221989207799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,10240,0.027702222267786663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,3584,0.054951111475626625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,8192,0.035381333695517644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,12288,0.008031110796663497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,32,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,256,0.014851555228233337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,10240,0.06667644447750516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,2560,0.007430222299363878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,4096,0.007631111476156447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,8192,0.03769066598680284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,3584,0.011004444625642566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,256,0.0028284444577164124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,5120,0.030028445853127375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,512,0.0038622220357259116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,3584,0.010072000324726105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,8192,0.02072177827358246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,64,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,7168,0.010055999788973067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,10240,0.012076444096035428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,6144,0.008987555901209513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,128,0.006304889089531369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,2048,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,1536,0.01146044416560067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,2048,0.006671111202902264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,128,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,1536,0.022437334060668945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,768,0.03123733401298523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,1024,0.275489780637953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,128,0.006625777731339137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,3072,0.01683999929163191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,128,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,12288,0.009217778013812171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,7168,0.016151999433835346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,2560,0.023746665981080797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,16384,0.026050666968027752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,256,0.016584888100624084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,128,0.030174222257402208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,16384,0.050620443291134305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,5120,0.008692444198661381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,2048,0.03525155451562669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,1536,0.026123555170165166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,5120,0.11440977785322402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,32,0.0018800000349680583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,7168,0.012055111428101858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,10240,0.03924622138341268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,3072,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,7168,0.01817244456874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,1536,0.009314667019579146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,5120,0.13148444228702122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,512,0.005759999983840519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,768,0.01767200065983666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,12288,0.013449778159459433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,512,0.02437422176202138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,2048,0.047769778304629855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,2048,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,2560,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,3584,0.014668444792429606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,128,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,65536,0.08280355400509305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,4096,0.008400888906584846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,7168,0.019946666227446664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,128,0.0052879999081293745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,1024,0.012596444123321109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,3584,0.027686221732033625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,3072,0.02064533366097344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,5120,0.22202489111158583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,16384,1.9736845228407118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,3584,0.02824266751607259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,10240,0.018191110756662156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,7168,0.01643111142847273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,2048,0.009119110802809397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,65536,0.06623822450637817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,256,0.008359111017651027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,2048,0.03791111045413547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,32,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,1536,0.0086666668454806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,1536,0.029905776182810467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,1024,0.0041999999019834726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,5120,0.024698666400379602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,1536,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,1536,0.023823110593689814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,65536,0.12869600454966226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,1024,0.023756444454193115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,2560,0.006956444846259222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,65536,0.3077404499053955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,32,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,6144,0.009768889182143742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,6144,0.03975111246109009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,1024,0.02619733413060506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,128,0.006923555499977536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,5120,0.02723466687732273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,5120,0.009111999637550777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,768,0.0031991110493739447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,7168,0.06952177815967135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,4096,0.08915644221835667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,768,0.0033084443873829315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,1536,0.042266666889190674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,256,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,32,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,128,0.0018657777044508194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,12288,0.018216000662909616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,12288,0.015789333317014907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,65536,0.1309093369377984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,4096,0.024697777297761705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,8192,0.029712888929578993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,7168,0.051721778180864125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,65536,0.17485777537027994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,6144,0.016891555653678048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,4096,0.007908444437715743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,6144,0.03422133458985223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,16384,0.03109333250257704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,768,0.007000889215204451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,1536,0.02676711148685879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,768,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,8192,0.0184879998366038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,12288,0.01923466722170512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,768,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,128,0.004249777644872665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,3072,0.02404533326625824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,2048,0.07151822249094646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,8192,0.01276266657643848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,65536,0.048264887597825795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,7168,0.03700266612900628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,6144,0.018191999859280057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,6144,0.03767911261982388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,1024,0.014456889695591398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,3584,0.013615111509958902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,2048,0.011725333001878528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,32,0.0070266665683852295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,1024,0.017889777819315594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,1024,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,64,0.0014808888857563336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,5120,0.008943999807039896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,128,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,256,0.015095111396577625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,6144,0.07487555344899495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,65536,0.14517245027754042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,16384,0.06137688954671224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,64,0.0018871112002266778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,16384,0.017154667112562392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,2560,0.018757333358128864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,768,0.02016266683737437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,1536,0.004404444247484207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,65536,0.03529777791765001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,512,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,8192,0.013753777576817406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,8192,0.021591110361946955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,7168,0.012256888879670037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,128,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,8192,0.013233777549531726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,2048,0.017867555220921833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,256,0.00217333332531982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,6144,0.014787554740905762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,7168,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,128,0.015795555379655626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,64,0.006635555376609166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,3072,0.009670221971140968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,5120,0.013040000365840064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,768,0.0025475554996066624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,2560,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,32,0.0018293333964215384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,10240,0.01769333415561252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,256,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,10240,0.030447999636332195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,32,0.0018657777044508194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,65536,0.13025866614447698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,16384,0.06116000148985121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,10240,0.015547555353906421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,6144,0.01866666641500261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,5120,0.009296000003814697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,3072,0.11994932757483588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,128,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,12288,0.2503697872161865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,4096,0.01236533373594284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,128,0.008965333302815756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,2560,0.014845333165592618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,4096,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,32,0.013799111048380533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,1536,0.009141333401203156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,4096,0.007720000214046902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,12288,0.2580782307518853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,128,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,128,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,1024,0.03641688823699951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,256,0.006177777631415262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,512,0.0038648889296584656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,4096,0.012381333443853589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,10240,0.14944622251722547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,8192,0.1256951093673706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,8192,0.018023111753993563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,3584,0.007640889121426477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,6144,0.026720000637902155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,1024,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,7168,0.00960177762640847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,512,0.0032106666929192017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,12288,0.017158221867349412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,65536,0.08818399906158447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,1024,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,3584,0.09115733040703668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,2560,0.16605510976579455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,8192,0.01074577785200543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,12288,0.013780444032616086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,65536,0.05108977688683403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,4096,0.024137778414620295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,128,0.014155555102560254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,4096,0.017660443981488545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,256,0.018834667073355783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,3072,0.006425777657164468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,7168,0.07700800233417086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,2048,0.01477599971824222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,2048,0.006849777781301075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,4096,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,65536,0.07430933581458198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,7168,0.013304000099500021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,512,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,3072,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,16384,0.04378755556212532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,1536,0.010386666489972008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,65536,0.09940977891286214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,3072,0.01311377767059538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,128,0.012884444660610624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,2560,0.013468444347381592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,16384,0.04231022132767571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,32,0.0022613333745135204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,1024,0.005622222191757626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,2560,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,8192,0.09257688787248398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,4096,0.009986666341622671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,3584,0.007633777956167857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,1536,0.0069013333155049225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,10240,0.02665866745842828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,8192,0.014450665977266101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,2560,0.010391111175219217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,512,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,6144,0.01998844411638048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,1536,0.10170666376749675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,3584,0.031119111511442397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,128,0.02029333346419864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,10240,0.02461599972512987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,2560,0.026878222823143005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,2048,0.00654755574133661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,12288,0.10209688875410293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,128,0.00830488900343577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,768,0.0035146665241983202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,64,0.0018133332745896445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,4096,0.04069777660899692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,2048,0.014020444618331062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,16384,0.03878755701912774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,3072,0.028184887435701158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,65536,0.1546880006790161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,65536,0.06989600261052449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,4096,0.11850310696495904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,65536,0.0663164456685384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,512,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,65536,0.10154222117529975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,6144,0.020604444874657523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,65536,3.1148355272081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,6144,0.026047110557556152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,7168,0.006912888752089606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,8192,0.020589333441522386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,3584,0.012878222597969903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,10240,0.2228399912516276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,512,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,2048,0.09412622451782227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,1536,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,5120,0.008976889153321585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,128,0.018127111925019156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,6144,0.015796444482273526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,16384,0.33484355608622235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,65536,0.09595377577675714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,3584,0.00832444429397583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,6144,0.025472000241279602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,512,0.008744888835483128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,16384,0.03347644540998671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,1024,0.005611555443869696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,3072,0.01445777714252472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,768,0.02288355595535702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,65536,0.5511866675482856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,5120,0.010695111420419482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,6144,0.32750844955444336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,65536,0.090229332447052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,16384,0.026138666603300307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,512,0.027048889133665297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,768,0.004523555437723796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,12288,0.01850311126973894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,16384,0.03790400094456143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,8192,0.011074666347768573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,65536,1.103485319349501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,32,0.014879999889267815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,12288,0.031515555249320135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,12288,0.04676711228158739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,12288,0.005170666509204441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,3072,0.008090666598743862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,8192,0.012719111310111152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,7168,0.02879022227393256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,4096,0.011032889286677042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,1536,0.01924800044960446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,768,0.01682844426896837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,2048,0.016184889607959323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,4096,0.012325333224402534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,128,0.0325564444065094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,4096,0.029041777054468792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,3072,0.00723555518521203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,5120,0.027650665905740526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,2560,0.007367110914654202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,16384,0.017851556340853374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,256,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,3072,0.01521511044767168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,10240,0.02665866745842828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,1536,0.018407111366589863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,65536,0.05851378043492635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,7168,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,512,0.0024862223201327855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,2560,0.14365243911743164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,7168,0.04266666703754001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,64,0.0017964444640609955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,3072,0.1053324474228753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,128,0.0024622221373849446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,12288,0.029771556456883747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,10240,0.01443555619981554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,128,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,5120,0.009022222624884712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,6144,0.021967111362351313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,3072,0.004551110996140374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,512,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,2560,0.18470222420162627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,16384,0.02497244377930959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,6144,0.016155555844306946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,128,0.010374222364690568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,2048,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,12288,0.024715556038750544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,256,0.023065777288542852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,1024,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,2560,0.018619555566045973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,128,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,128,0.014465777410401238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,6144,0.009655999640623728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,32,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,3584,0.08884621991051568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,8192,0.03834399912092421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,7168,0.019878221882714164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,2048,0.00602311102880372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,1536,0.023736000061035156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,5120,0.05787999762429131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,256,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,256,0.005601777798599667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,7168,0.008813333180215623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,2048,0.011972444752852121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,1024,0.005839111076460944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,4096,0.032708442873424955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,32,0.0047866664826869965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,1536,0.012459555433856117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,768,0.0165964447789722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,12288,0.017768889665603638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,8192,0.030454221698972914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,2560,0.04086755712827047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,6144,0.019139554765489366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,768,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,5120,0.021909332937664453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,1024,0.01480977733929952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,16384,0.20112178060743544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,7168,0.02186577849917942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,4096,0.015538665983412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,2560,0.024736889534526404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,1024,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,1536,0.04121422105365329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,12288,0.028881778319676716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,256,0.0021546667234765156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,256,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,3072,0.024346666203604803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,8192,0.12706666522555882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,4096,0.007993777592976889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,7168,0.01891377733813392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,512,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,7168,0.023217777411142986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,7168,0.021771555145581562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,256,0.006606222026877933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,2048,0.0140346669488483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,12288,0.02402399977048238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,4096,0.02323466704951392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,8192,0.015411555767059326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,16384,0.06374577681223552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,64,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,256,0.002493333278430833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,256,0.014785778191354541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,12288,0.06764088736640082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,16384,0.03130044539769491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,5120,0.03325600094265408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,8192,0.0803697771496243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,12288,0.03707289033465915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,16384,0.08742666906780666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,3072,0.00554044461912579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,32,0.001842666624320878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,256,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,3584,0.005879999862776862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,12288,0.014459555347760519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,8192,0.011044444309340583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,3584,0.007442666424645319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,3072,0.012958221965365939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,256,0.0028604444944196274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,64,0.002485333424475458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,256,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,2048,0.025786666406525507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,2560,0.023435556226306494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,128,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,128,0.014825777875052558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,65536,0.17692177825503877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,256,0.02159999973244137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,2048,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,256,0.006311999840868845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,12288,0.07728622357050578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,4096,0.009686222506894005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,4096,0.009665778113736046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,2560,0.006911110960774952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,7168,0.040948443942599826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,5120,0.24527555041842988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,3584,0.2924773428175184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,10240,0.014137778017255994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,6144,0.1193288829591539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,4096,0.014404444230927361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,7168,0.03150044547186957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,512,0.02786133355564541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,3584,0.004571555389298333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,65536,0.1211048894458347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,3072,0.007223111059930589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,512,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,16384,0.020791111720932853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,512,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,1536,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,2048,0.010981332924630908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,3072,0.24955378638373482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,3584,0.022660444180170696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,1536,0.00814933329820633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,8192,0.06335910823610094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,65536,0.011357333097192975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,768,0.011357333097192975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,65536,0.35038667254977757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,2560,0.008275555239783393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,1024,0.011028444601429833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,8192,0.08685421943664551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,4096,0.03503288825352987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,32,0.03591377867592706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,8192,0.016156444946924847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,3072,0.005873777800136142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,4096,0.007814221911960179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,4096,0.016936888297398884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,7168,0.10275644726223415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,3072,0.11911911434597439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,2560,0.01854222185081906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,65536,0.1485511064529419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,7168,0.0432168874475691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,4096,0.04964888758129544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,7168,0.038557334078682795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,65536,0.12605510817633733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,8192,0.010384889112578498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,5120,0.02610933283964793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,12288,0.04384444488419426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,7168,0.10258044136895074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,1536,0.010984888507260216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,3584,0.048764444059795804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,3072,0.020245333512624104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,10240,0.027190221680535212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,1536,0.00628977765639623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,7168,0.008961777720186446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,5120,0.019596444235907663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,12288,0.014241778188281588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,1536,0.01850133306450314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,4096,0.006226666685607698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,12288,0.04100799891683791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,512,0.0028719999310043124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,128,0.018423999349276226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,6144,0.024079999989933435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,7168,0.029679109652837116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,1024,0.008977777428097194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,6144,0.06796533531612821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,65536,0.16708266735076904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,16384,0.022022222479184467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,65536,0.06461511055628459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,256,0.0022026666750510535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,3584,0.01126488877667321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,4096,0.1542417738172743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,512,0.005982222242487802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,6144,0.009663999908500248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,1536,0.03148355417781406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,16384,0.01512977812025282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,7168,0.013931555880440606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,2560,0.015493333339691162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,12288,0.00720177756415473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,8192,0.0465848876370324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,256,0.01516266663869222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,65536,0.18681332800123426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,2560,0.05747822258207533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,256,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,7168,0.09983377986484104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,4096,0.06681422392527263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,5120,0.01585244470172458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,5120,0.009458666874302758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,3584,0.020813332663642038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,2560,0.008354666332403818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,7168,0.02331022255950504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,8192,0.048135999176237315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,3584,0.011381333072980246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,1024,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,512,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,16384,0.04489333430926005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,3072,0.016185777054892648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,256,0.004917333523432414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,5120,0.03446311089727614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,12288,0.013059555656380124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,8192,0.03618311219745212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,4096,0.008286221987671321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,128,0.01514400045077006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,2048,0.008004444340864817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,5120,0.012897777888509961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,32,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,1024,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,512,0.02197244432237413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,6144,0.1516640053855048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,512,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,3072,0.022963555322753057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,3072,0.011309333145618439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,2560,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,256,0.01484444406297472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,512,0.002487999904486868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,16384,0.47642935646904844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,4096,0.01889866590499878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,7168,0.013580444786283704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,2048,0.00797333319981893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,8192,0.023949333363109167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,12288,0.12823200225830078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,1536,0.02281422250800663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,3584,0.010469333165221743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,7168,0.009855111440022787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,5120,0.01279022213485506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,6144,0.04254666633076138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,12288,0.27458755175272626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,65536,0.28795289993286133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,256,0.005356444252861871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,3072,0.007592000067234039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,1024,0.012744888663291931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,1536,0.008271111382378472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,6144,0.009648000200589498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,1536,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,1024,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,2048,0.012738666600651212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,3072,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,3072,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,256,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,256,0.0025315555847353404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,32,0.0016791110651360618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,256,0.0032239999208185407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,5120,0.2801528771718343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,2560,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,7168,0.013223110801643796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,512,0.006974221931563483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,12288,0.03456711106830173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,6144,0.008608889248636033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,5120,0.02017600006527371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,256,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,65536,0.06324533621470134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,2048,0.02914666798379686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,3584,0.0266311119000117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,7168,0.045908444457583956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,3072,0.016763556334707472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,1536,0.006903111106819577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,3584,0.009669333696365356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,10240,0.015425778097576566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,3072,0.015511999527613321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,5120,0.10796178049511379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,7168,0.011408000356621213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,6144,0.06159733401404487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,16384,0.03546488947338528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,65536,0.05207822057935926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,3072,0.027710222535663184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,16384,0.02684177789423201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,12288,0.052967998716566295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,3584,2.643562740749783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,3584,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,6144,0.027031110392676458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,3072,0.03613244493802389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,8192,0.02092711130777995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,16384,0.06888177659776476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,12288,0.025016888976097107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,10240,0.03188177943229675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,2048,0.006986666884687211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,256,0.013401777380042605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,256,0.007685333490371704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,512,0.0052871112194326185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,65536,0.19311911529964873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,2048,0.008310221963458592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,4096,0.031103110975689355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,10240,0.035206221871905856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,1536,0.01847466660870446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,7168,0.011015999648306104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,256,0.021606221795082092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,64,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,16384,0.07795466979344685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,1024,0.06864088773727417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,65536,0.12742933962080213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,16384,0.07866133583916558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,512,0.004121777911980947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,12288,0.033655110332700945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,768,0.0029324444217814338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,1024,0.011992000043392181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,8192,0.0107004443804423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,3072,0.014825777875052558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,64,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,5120,0.016214221715927124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,3072,0.03125155634350247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,12288,0.022634666826989915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,1536,0.03199199835459391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,6144,0.018244443668259513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,32,0.006765333314736684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,5120,0.007272000114123027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,3584,0.013442666994200813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,512,0.008638222184446123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,512,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,2048,0.03476266728507148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,2560,0.020616888999938965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,2560,0.01888088881969452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,128,0.0052897776994440295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,1024,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,8192,0.011064888702498542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,768,0.006970666348934174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,3584,0.014102222190962898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,3584,0.014719110396173267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,32,0.006224000205596288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,64,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,4096,0.1068835523393419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,768,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,768,0.013639111485746173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,3584,0.011701333026091257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,1024,0.007662222617202335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,8192,0.013304000099500021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,8192,0.011384000380833944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,7168,0.041379554404152766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,2560,0.007253333098358578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,1536,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,256,0.026053332620196875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,8192,0.022241777843899194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,65536,0.0880933337741428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,3072,0.004250666747490565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,128,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,3072,0.009454222189055549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,4096,0.04796355631616381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,16384,0.016152888536453247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,6144,0.014390221900410123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,8192,0.02092355489730835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,256,0.0025155554629034465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,16384,0.047601779301961265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,5120,0.008940444224410588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,10240,0.021264889174037512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,3584,0.02864977717399597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,7168,0.02314222190115187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,5120,0.004888888862397936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,768,0.015498666299713982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,512,0.17497511704762778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,7168,0.042191998826132886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,65536,0.0373324453830719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,2048,0.015511999527613321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,64,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,2048,0.00701155596309238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,512,0.014440000057220459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,3584,0.010051555931568146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,12288,0.04199022385809156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,8192,0.14652889304690891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,768,0.006045333213276333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,32,0.006673777682913675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,65536,0.06863200002246432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,256,0.02498222225242191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,1024,0.010017777482668558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,5120,0.01479466590616438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,16384,0.031898667414983116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,4096,0.018233777748213876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,256,0.04567555586496989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,10240,0.016235555211702984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,16384,0.029824889368481104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,512,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,8192,3.2477715810139975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,768,0.022410665949185688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,1024,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,1536,0.01830933325820499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,2048,0.012136889000733694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,768,0.03675377700063918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,1024,0.043713778257369995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,256,0.013442666994200813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,128,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,1024,0.006276444428496891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,2560,0.007321777443091075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,16384,0.059198220570882164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,6144,0.006158222340875202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,10240,0.03231822119818793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,7168,0.010398222340477837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,12288,0.023649776975313824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,3072,0.008151999778217739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,3072,0.01719466679626041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,768,0.010030222435792288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,2048,0.006660444455014334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,16384,0.017457778255144756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,16384,0.13795822196536595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,1024,0.006073777874310811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,8192,0.03731733229425218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,256,0.0063075555695427795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,1536,0.020240000552601285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,256,0.0027546667390399506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,1024,0.015026667051845126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,2560,0.00759733302725686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,12288,0.013391111459996967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,4096,0.055883553293016225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,10240,0.018600889378123812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,10240,0.28417245546976727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,2560,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,1024,0.02530488868554433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,3072,0.012442666623327466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,128,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,16384,0.024637333220905725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,1536,0.01924000018172794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,256,0.024396444360415142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,12288,0.01349511080318027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,6144,0.0195795562532213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,16384,0.14301510651906332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,7168,0.011471110913488599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,768,0.01885511146651374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,1024,0.0031964443624019623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,2560,0.020938666330443487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,1024,0.0034799998005231223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,512,0.0027804445061418745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,512,0.018982221682866413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,6144,0.004264889078007804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,1536,0.037187556425730385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,6144,0.23231556680467394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,16384,0.013853333062595792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,3072,0.00903555585278405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,1024,0.0063528890411059065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,3072,0.022967999180157978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,3584,0.014447111222479077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,16384,0.06081333425309923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,2560,0.11764978037940131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,2048,0.005396444350481033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,256,0.012120000190205045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,16384,0.08651822143130833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,1536,0.0036799998746977914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,64,0.0018657777044508194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,2048,0.00828888929552502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,7168,0.04043644335534837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,10240,0.013637333280510373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,8192,0.010495111346244812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,10240,0.015273777974976433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,512,0.0028577778074476454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,128,0.004889777965015835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,1024,0.0314319994714525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,768,0.019558222757445443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,768,0.0028373334142896864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,256,0.011030221978823343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,2560,0.007610666255156199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,2560,0.007601777712504069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,4096,0.04169866773817274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,128,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,8192,0.017937777770890128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,2560,0.014974221587181091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,16384,0.0429004430770874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,512,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,10240,0.15761600600348577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,1024,0.006248888870080312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,1536,0.009450666606426239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,65536,0.17285333739386666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,768,0.02751644452412923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,3072,0.009686222506894005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,4096,0.00832177781396442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,5120,0.005786666853560342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,12288,0.013451555536852943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,7168,0.013740444348918067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,4096,0.019511999355422128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,768,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,256,0.0038444445365005066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,256,0.0053093334039052325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,32,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,12288,0.033779554896884494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,2560,0.006245333287451003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,1024,0.004816888935036129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,32,0.00184888889392217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,512,0.0031591111587153543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,1536,0.005416000054942236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,5120,0.010086221827401055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,7168,0.01036533299419615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,12288,0.10576711098353068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,32,0.005207111024194294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,4096,0.16185243924458823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,32,0.002141333288616604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,1536,0.003469333259595765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,7168,0.07455466853247748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,1536,0.040728890233569674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,512,0.0025031111306614345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,12288,0.16881688435872397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,1024,0.03198044498761495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,1024,0.0031991110493739447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,10240,0.011751111182901593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,1536,0.004204444587230682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,3584,0.010453333457310995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,12288,0.041737778319252856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,1536,0.0042053332759274375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,5120,0.009161777794361115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,128,0.0024666666156715816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,4096,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,3072,0.019644444187482197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,2048,0.00453244439429707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,12288,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,12288,0.12835110558403864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,1536,0.003656889001528422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,5120,0.05308799942334493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,3584,0.15411999490525988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,3584,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,3584,0.007620444728268518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,2560,0.006648889018429651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,2048,0.004503111044565837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,2560,0.018545778261290658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,2560,0.010235555469989777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,128,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,32,0.023775110642115276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,512,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,3584,0.19296888510386148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,5120,0.05758044454786512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,3584,0.007324444750944774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,5120,0.00869777798652649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,4096,0.01273333364062839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,2560,0.011697777443461947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,1024,0.008300444318188561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,2048,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,65536,0.0956195526652866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,7168,0.17045511139763725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,256,0.0032471112079090546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,3584,0.008007110820876228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,1024,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,8192,0.022871999277008906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,768,0.0040960001448790235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,8192,0.06198933389451769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,2560,0.007886221839321984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,1536,0.015965332587560017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,1536,0.007984889050324758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,1536,0.02160800000031789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,2048,0.004596444467703502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,768,0.0038559999730851916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,128,0.022806222240130108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,5120,0.012235555383894177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,16384,0.03116977877087063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,2560,0.017398221625222098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,768,0.011692444483439127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,1024,0.0064586665895250105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,10240,0.035637332333458796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,768,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,8192,0.0225262227985594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,5120,0.019401777121755812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,5120,0.013789333403110504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,2560,0.016751999656359356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,10240,0.012072000238630505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,65536,0.10988089111116196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,512,0.012054222325483957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,2560,0.007666666474607255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,3584,0.02306933369901445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,16384,0.2221697833802965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,512,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,768,0.028352889749738906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,3072,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,32,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,5120,0.009524444739023844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,8192,0.010547555155224271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,10240,0.03801777627733018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,8192,0.00831822223133511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,2048,0.04832977718777127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,2048,0.010712888505723743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,256,0.008616888688670265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,128,0.005429333282841577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,1024,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,12288,0.02783999840418498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,1024,0.025182222326596577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,4096,0.014670222997665405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,2048,0.014428445034556918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,32,0.0018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,128,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,128,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,32,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,4096,0.11325600412156846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,6144,0.024367999699380662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,8192,0.24019734064737955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,2048,0.008277333445019191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,16384,0.05912799967659844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,256,0.0028071111689011254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,1024,0.03036622206370036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,1024,0.007074666519959767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,512,0.013883555928866068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,65536,0.08849155240588719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,1536,0.012671111358536614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,10240,0.04827022221353319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,7168,0.012776888906955719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,8192,0.011071999867757162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,16384,0.04893689023123848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,32,0.0019528888579871918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,12288,0.33631467819213867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,768,0.007333333293596904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,8192,0.014548444085650973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,3072,0.007691555553012424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,512,0.002720000015364753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,768,0.006304889089531369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,768,0.017839999662505258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,3584,0.029125332832336426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,768,0.0079297779334916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,768,0.02176711128817664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,7168,0.028152889675564233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,3584,0.009391999906963771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,2560,0.008664888640244802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,12288,0.0614195532268948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,16384,0.06973777876959907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,65536,0.05460000038146973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,2048,0.007054222126801808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,6144,0.044201778040991895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,7168,0.0052639999323421055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,512,0.002479111154874166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,12288,0.8301991356743706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,2048,0.00870488915178511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,8192,0.03538666831122504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,8192,0.03143111202451918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,2048,0.017232000827789307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,3584,0.06264177958170573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,32,0.0057653333577844836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,5120,0.0151555554734336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,1024,0.017088888419999015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,1536,0.0185253338681327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,3072,0.022231111923853557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,4096,0.01146399974822998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,65536,0.06440444125069512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,2560,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,10240,0.004907555464241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,65536,0.1406195561091105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,768,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,3072,0.007430222299363878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,65536,0.2861768934461805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,2048,0.16135111120012072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,128,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,8192,0.15621066093444824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,256,0.002785777880085839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,12288,0.016846223009957206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,3072,0.023919110496838886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,512,0.003170666595300039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,65536,1.1648427115546334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,7168,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,4096,0.01349066694577535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,6144,0.03241689006487528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,65536,0.5625128746032715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,2048,0.04934044347869026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,512,0.0063546668324205614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,512,0.005944889038801193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,5120,0.01089599976936976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,512,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,10240,0.03405333227581448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,768,0.0064088888466358185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,6144,0.009288888838556077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,1536,0.003548444559176763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,128,0.0018213333355055915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,12288,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,6144,0.4118391142951117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,2560,0.010472888747851053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,6144,0.011012444065676795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,65536,0.20120267073313394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,3072,0.0148026661740409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,12288,0.016543999314308167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,64,0.003144888828198115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,2048,0.0396035545402103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,7168,0.022375999225510493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,6144,0.034918222162458636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,2560,0.0151573336786694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,65536,0.14740178320142958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,2560,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,1024,0.017850667238235474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,16384,0.1753235525555081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,2048,0.00673333348499404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,256,0.015273777974976433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,7168,0.038224889172448054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,768,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,5120,0.011385777758227454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,5120,0.033434665865368314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,8192,0.07794488800896539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,128,0.02107644412252638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,65536,0.06166311105092367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,10240,0.01670222149954902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,128,0.002136888810329967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,4096,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,12288,0.05422755413585239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,5120,0.014890667464998035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,2560,0.0140328887436125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,512,0.005447111195988125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,16384,0.17769422796037462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,10240,0.01910666624704997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,12288,0.025013332565625507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,2048,0.004426666845877965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,512,0.027122666438420612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,16384,0.053800887531704374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,5120,0.008762666748629676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,10240,0.02513066596455044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,8192,0.030821333328882854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,256,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,512,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,512,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,16384,0.03711644477314419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,65536,0.011709333293967776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,16384,0.0677857796351115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,512,0.002509333400262727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,4096,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,768,0.0031742221779293488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,16384,0.03209066722128127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,5120,0.008609777523411645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,768,0.003192000091075897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,5120,0.2502640088399251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,2048,0.01588800052801768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,3072,0.01738577749994066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,3072,0.036333332459131874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,4096,0.027618666489919026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,3072,0.008319999608728621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,4096,0.025641777449183997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,65536,0.00656355544924736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,65536,0.011020444333553314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,1536,0.03645244571897719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,256,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,128,0.005277333160241445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,1024,0.019239111079110038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,512,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,1536,0.38198312123616535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,7168,0.010449777874681683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,16384,0.01718755563100179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,128,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,8192,0.01806311143769158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,3584,0.08287377489937676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,6144,0.1765164401796129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,65536,0.6250702010260688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,768,0.009582222335868413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,65536,0.2026684416664971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,3584,0.006246222390068903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,4096,0.00795111142926746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,1024,0.006021333237489064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,32,0.0020862221717834473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,12288,0.2926657729678684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,6144,0.030569778548346624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,256,0.003660444584157732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,8192,0.018336888816621568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,16384,0.015848888291252982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,6144,0.009334222310119206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,2560,0.05119644602139791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,32,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,1536,0.006620444357395172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,3584,0.18452978134155273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,6144,0.009943111075295342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,5120,0.11335199409061009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,12288,0.030576888057920668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,8192,0.020623111062579684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,2560,0.007321777443091075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,12288,0.017415111263593037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,256,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,512,0.002814222127199173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,3072,0.018527110417683918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,2560,0.01311377767059538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,12288,0.01645511057641771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,8192,0.02016622159216139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,256,0.015311110350820752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,5120,0.013210666676362356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,65536,0.14195556110805935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,512,0.014828445182906257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,12288,0.04982044299443563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,10240,0.03843466771973504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,32,0.011424889167149862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,3584,0.018639999959203932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,3072,0.007583111524581909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,6144,0.1832737790213691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,5120,0.01942044496536255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,64,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,768,0.0031831111345026228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,1536,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,5120,0.02865866820017497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,512,0.0028275555620590844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,3072,0.013088888592190213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,16384,0.022389334109094407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,65536,0.6379297574361166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,3584,0.008904889225959778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,1536,0.006559111177921295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,2048,0.005532444351249271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,768,0.015164444843928019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,256,0.01578044394652049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,6144,0.04871199859513176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,6144,0.03497688968976339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,7168,0.010024000373151567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,3584,0.009213333328564962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,12288,0.48629601796468097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,3584,0.02252799934811062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,512,0.007703111403518253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,32,0.0069875551594628235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,7168,0.06370933188332452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,128,0.01478666729397244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,7168,0.13226221667395696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,16384,0.03996266590224372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,4096,0.03448444604873657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,7168,0.018034666776657104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,7168,0.022276444567574397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,3072,0.00848444468445248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,4096,0.011419555379284753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,4096,0.04884799983766344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,3584,0.023766222927305434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,256,0.02269688910908169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,1024,0.007268444531493717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,6144,0.004934222333961063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,4096,0.09035378032260472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,2560,0.019260444574885897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,1024,0.04203377829657661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,8192,0.10430400239096747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,8192,0.011058666639857821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,256,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,16384,0.05686933464474148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,16384,0.016121778223249648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,16384,0.2494871086544461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,1024,0.007962666451931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,3072,0.01036000003417333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,12288,0.05265066689915127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,64,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,256,0.01919822229279412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,128,0.004908444566859139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,6144,0.10692889160580105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,3584,0.04242222176657783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,2048,0.015097777048746744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,512,0.0028248888750871024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,512,0.007317333585686154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,512,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,3584,0.024604444702466328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,7168,0.011742221812407175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,512,0.006414222220579783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,1536,0.003863999827040566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,256,0.018195556269751657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,6144,0.00978133330742518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,3584,0.05866489145490858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,16384,0.07414399915271334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,3072,0.01793066660563151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,1536,0.006187555690606435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,512,0.015507555670208402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,768,0.00793422261873881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,2048,0.07644710938135783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,512,0.005373333477311664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,512,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,1024,0.003220444545149803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,3584,0.01179377817445331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,8192,0.02401599950260586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,2048,0.02922399838765462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,4096,0.14689511722988552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,128,0.027411555250485737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,2560,0.04862666792339749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,4096,0.008080888953473832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,2560,0.008832000195980072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,2048,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,128,0.0058204444746176405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,128,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,2560,0.009722666607962714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,5120,0.41354044278462726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,512,0.0025946667624844448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,1024,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,7168,0.010384000009960597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,10240,0.022652443912294176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,768,0.002576888849337896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,12288,0.025399999486075506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,1024,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,3584,0.007605333295133378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,12288,0.2074364423751831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,256,0.015354666444990369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,64,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,1024,0.004537777768241035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,1024,0.006312888943486744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,2560,0.020603555772039626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,4096,0.011074666347768573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,8192,0.01274400038851632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,768,0.01831022236082289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,256,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,3584,0.01800622211562263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,256,0.017289777596791584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,2048,0.007992000215583378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,10240,0.01739644507567088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,1536,0.004299555387761858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,8192,0.005578666511509154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,768,0.0031822222388452957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,12288,0.02817866537306044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,10240,0.01198755535814497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,128,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,7168,0.01990133358372582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,12288,0.018716444571812946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,512,0.006956444846259222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,12288,0.210589329401652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,8192,0.030305776331159804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,512,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,3584,0.01404711107412974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,16384,0.14653955565558538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,16384,0.053820444477929004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,4096,0.008638222184446123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,768,0.017835555805100333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,65536,0.06590933269924588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,256,0.016533333394262526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,7168,0.007159999675220913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,1024,0.033590220742755465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,1024,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,4096,0.019558222757445443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,128,0.0028248888750871024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,256,0.014817777607176038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,2048,0.023312888211674158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,768,0.004911111046870549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,256,0.0024888888001441956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,5120,0.01888177792231242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,2560,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,32,0.005972444597217772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,7168,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,256,0.01809599995613098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,8192,0.07835555738872953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,4096,0.0876693328221639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,12288,0.019824888971116807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,7168,0.011417778001891242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,128,0.002495999965402815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,3072,0.014087110757827759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,6144,0.013468444347381592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,128,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,6144,0.028555555476082697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,6144,0.011057777537239922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,12288,0.04403466648525662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,12288,0.053073777092827686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,32,0.013092444174819522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,10240,0.029451555675930444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,6144,0.009659555223253038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,10240,0.01862133377128177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,8192,0.18700088395012748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,2048,0.01608088943693373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,16384,0.021850667066044275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,1536,0.0063582224150498705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,1536,0.00629333323902554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,4096,1.833446290757921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,12288,0.06109155548943413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,7168,0.03462933169470893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,12288,0.018515555395020377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,32,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,1536,0.0038542221817705366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,5120,0.008952000074916417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,6144,0.02294755478700002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,1024,0.02144977781507704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,16384,0.0359866652223799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,128,0.0021271111650599376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,4096,0.009682666924264696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,1536,0.023747555083698694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,64,0.025906667113304138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,5120,0.004880889008442561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,32,0.001785777716173066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,16384,0.04337777694066366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,65536,0.10916088687049018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,6144,0.011557333171367645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,256,0.003132444495956103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,32,0.0076453329788313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,4096,0.00795733349190818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,4096,0.016556445095274184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,3584,0.05774311224619547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,64,0.007635555333561367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,7168,0.022665777140193518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,256,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,64,0.0032106666929192017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,4096,0.007978666987684038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,7168,0.010701333483060202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,3584,0.02278577784697215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,32,0.0021404443929592767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,6144,0.0388995541466607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,8192,0.02607822252644433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,3072,0.0059164443777667145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,1536,0.00851822230550978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,128,0.014471999473041959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,5120,0.009335999687512716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,128,0.0031315556002987754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,256,0.01718577742576599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,3072,0.03600177831119961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,256,0.006165333506133821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,7168,0.009995555712117089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,10240,0.028493334849675495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,256,0.005632888939645555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,256,0.0069066666894488865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,16384,0.02247200078434414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,16384,0.016511999898486666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,64,0.020042666130595736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,128,0.002171555534005165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,12288,0.10456355412801106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,6144,0.01680177781316969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,5120,0.05015733175807529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,10240,0.015823110938072205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,768,0.0029120000286234748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,1024,0.04455644554562039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,6144,0.13364444838629827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,2048,0.004564444637960858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,1024,0.023443554838498432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,768,0.002567111204067866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,64,0.0018613334331247543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,768,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,768,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,256,0.012605333493815528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,16384,0.08864088853200276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,65536,0.010363555616802638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,2048,0.02084355552991231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,8192,0.01330844478474723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,1536,0.009914666414260864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,16384,1.6565430959065754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,12288,0.04419555597835117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,2048,0.008185777399275038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,5120,0.04291466540760464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,768,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,3072,0.005865777946180767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,65536,0.14318488703833684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,10240,0.014540443817774454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,1536,0.025995555851194594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,8192,0.033124443557527326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,16384,0.015839111473825242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,7168,0.01444355481200748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,128,0.02091911103990343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,10240,0.04389155573315091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,12288,0.018224888377719455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,12288,0.09437066978878444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,1024,0.01344622257683012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,8192,0.06016533242331611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,5120,0.14421778255038792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,6144,0.019861333900027804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,512,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,2560,0.008015111088752747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,3584,0.026188444760110643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,12288,0.0052693333062860705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,2560,0.014241778188281588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,128,0.0028435554769304064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,6144,0.016895999511082966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,1024,0.027475555737813313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,4096,0.7257155312432183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,4096,0.022312000393867493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,2048,0.027678221464157104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,65536,0.06847200128767225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,128,0.013216888739003075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,3072,0.012064889073371887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,65536,0.19738488727145723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,64,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,7168,0.013434666726324292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,12288,0.014793778459231058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,16384,0.01461155547036065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,32,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,64,0.0018577778504954444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,3584,0.010334221853150262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,256,0.011026666396194033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,12288,0.026962666047943964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,8192,0.02386666668785943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,6144,0.009318222602208456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,3584,0.039529777235454984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,512,0.011740444435013665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,3072,0.007313778003056844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,3584,0.010218666659461128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,512,0.0052782222628593445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,12288,0.053525331947538585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,2048,0.030920889642503526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,16384,0.06739911105897692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,2560,0.021317332983016968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,2560,0.007114667031500075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,3072,0.007458666960398356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,8192,0.011025778121418424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,5120,0.10894399881362915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,7168,0.030748443471060857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,2560,0.06836977932188246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,3072,0.016194666425387066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,32,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,3584,0.008042666647169325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,65536,0.14238488674163818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,1024,0.02045866681469811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,1024,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,32,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,10240,0.017614222235149808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,2048,0.015792888071801927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,32,0.005274666680230035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,512,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,8192,0.015104888214005364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,1024,0.00350577798154619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,1024,0.006042666733264923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,1536,0.0034968890249729156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,2560,0.00980711148844825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,2048,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,6144,0.03410222133000692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,768,0.009373333719041612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,256,0.01718311177359687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,7168,0.018221333622932434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,4096,0.013048888908492194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,7168,0.02362844513522254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,1024,0.017494221528371174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,4096,0.01218933363755544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,4096,0.0075786668393347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,6144,0.01144444445768992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,512,0.00831200016869439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,3072,0.012176888684431711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,3584,0.07012799713346693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,6144,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,16384,0.016522667474216886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,5120,0.02404088940885332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,4096,0.040848887628979154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,8192,0.011172444456153445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,3072,0.01642311116059621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,128,0.010122666756312052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,16384,0.05408089028464424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,3584,0.010728889041476779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,7168,0.1442399952146742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,256,0.03712888889842563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,2048,0.008360889222886827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,1024,0.015120888749758402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,64,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,32,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,3072,0.0069022224181228215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,6144,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,128,0.005344888816277186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,256,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,12288,0.10561244355307685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,16384,0.024335111180941265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,2560,0.00738844441043006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,7168,0.012076444096035428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,256,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,6144,0.04717155628734165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,8192,0.009822222093741098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,1024,0.006597333484225803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,10240,0.016936888297398884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,7168,0.03975199990802341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,16384,0.016148444679048326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,64,0.0018568889548381169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,256,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,65536,0.10506133238474528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,2560,0.2909235424465603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,8192,3.6938069661458335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,768,0.020991999242040846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,16384,0.02313066687848833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,10240,0.029331554969151814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,1024,0.061871998839908175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,64,0.008928000099129146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,10240,0.02180444366402096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,3584,0.0207360006040997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,128,0.008025777836640676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,10240,0.010268444816271463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,8192,0.051810665263069995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,65536,0.14438400003645155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,8192,0.02514311174551646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,5120,0.011125333607196808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,768,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,2560,0.011585777832402123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,6144,0.017786666750907898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,16384,0.021314667330847845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,768,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,7168,0.06243466668658786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,3072,0.009711110757456886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,2560,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,3584,0.024143111374643113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,5120,0.15241866641574436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,1536,0.0064284445510970215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,4096,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,4096,0.008772444393899705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,64,0.0022008888837363985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,64,0.0018435555199782054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,8192,0.017484444710943434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,3584,0.01183733344078064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,2560,0.016574222180578444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,4096,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,3072,0.011001778145631155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,4096,0.011006222003036074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,2560,0.011357333097192975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,32,0.005958222266700532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,768,0.002879111096262932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,6144,0.01681688924630483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,64,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,3584,0.024897777371936377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,3584,0.004943111290534337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,3584,0.011074666347768573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,1536,0.03281777766015794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,768,0.008669333325492011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,32,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,16384,0.10975288682513767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,16384,0.044581333796183266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,1024,0.006261333409282897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,2560,0.011864888999197217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,7168,0.13414311408996582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,2048,0.0069324444565508105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,12288,0.46465153164333767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,12288,0.03345422281159295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,7168,0.04628355635537041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,3584,0.01310400002532535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,12288,0.02571022179391649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,256,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,65536,0.09641955296198527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,128,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,65536,0.06019733349482218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,2048,0.0063493334584765965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,65536,0.1894622246424357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,65536,0.1602835522757636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,4096,0.008069333102968004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,12288,0.08100533485412598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,7168,0.010764444039927589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,8192,0.13066311677296957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,16384,0.020601777566803824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,256,0.008743999732865227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,768,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,12288,0.026939556002616882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,3072,0.0058133333093590206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,7168,0.06694577799903022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,3072,0.008233778178691864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,7168,0.012631110846996307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,6144,0.03988533218701681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,8192,0.057335999276902944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,32,0.004039999925427967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,1024,0.004280000097221798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,3072,0.13387021753523085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,4096,0.04531288809246487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,1024,0.011312888728247749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,4096,0.02292888859907786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,3584,0.004547555413511065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,4096,0.0042035554846127825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,6144,0.009659555223253038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,512,0.005526222288608551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,7168,0.10266577535205418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,64,0.005220444252093633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,10240,0.14843289057413736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,10240,0.0319164428446028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,10240,0.16413954893747965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,5120,0.009334222310119206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,768,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,16384,0.034613334470325045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,7168,2.784796396891276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,512,0.005965333431959152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,2048,0.018555555078718398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,2560,0.00702311098575592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,1536,0.01683022247420417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,6144,0.011948444777064852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,8192,0.03203377789921231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,3072,0.004916444420814514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,3072,0.005344888816277186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,4096,0.1577448844909668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,65536,0.19237066639794242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,7168,0.06310044394599067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,8192,0.028984887732399836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,12288,0.01055111073785358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,256,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,32,0.0019351111518012153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,16384,0.0985564457045661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,3072,0.05786489115820991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,64,0.016550223032633465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,128,0.015448888142903646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,2560,0.09150666660732693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,1536,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,65536,0.2317670981089274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,8192,0.7250613106621636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,65536,0.1877217822604709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,5120,0.017490666773584154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,10240,0.023352000448438857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,8192,0.06913422213660346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,128,0.00638755535085996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,128,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,2560,0.3239324357774523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,1024,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,65536,0.17344444327884248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,768,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,3584,0.00775733341773351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,7168,0.00850755555762185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,32,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,10240,0.01844800015290578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,4096,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,3072,0.008044444852405125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,3072,0.008987555901209513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,32,0.0035111109415690103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,10240,0.03461688756942749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,2560,0.012656889028019376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,32,0.0020684444655974707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,10240,0.2757786644829644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,6144,0.056925336519877114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,7168,0.03327644533581204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,16384,0.02356622285313076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,12288,0.046032889021767505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,1024,0.006082666830884085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,2560,0.009669333696365356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,768,0.007821333077218797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,512,0.0069066666894488865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,6144,0.03154666556252374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,4096,0.017893332574102614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,8192,0.09579733345243667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,64,0.00656088896923595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,3584,0.008654221892356873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,512,0.029872000217437744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,4096,0.010867555936177572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,128,0.013471110827393003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,5120,0.10593510998619927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,16384,0.03723999857902527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,2048,0.11431022485097249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,10240,0.027473777532577515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,4096,0.007657777931955125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,2560,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,2048,0.008376888930797577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,1024,0.020978666014141504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,16384,0.029273778200149536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,128,0.015150222513410779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,512,0.009007111191749573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,8192,0.021721776988771226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,8192,0.03847733471128676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,16384,0.02384000023206075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,2560,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,65536,0.13838844829135472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,5120,0.006310222049554189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,2048,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,768,0.015680889288584392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,2048,0.010000000397364298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,128,0.014348444011476306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,12288,0.05482755435837639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,3584,0.018869333797030978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,4096,0.009626666704813639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,3072,0.005601777798599667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,512,0.011832888755533429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,1024,0.006508444332414203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,4096,0.011951111257076263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,1024,0.005974222388532426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,8192,0.025362667110231187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,2048,0.006934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,2560,0.11517866452534993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,5120,0.00812533332241906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,768,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,1536,0.005963555640644497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,5120,0.22930577066209581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,3072,0.00793599999613232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,128,0.0022035555707083806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,16384,0.06480622291564941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,3584,0.014641778336630927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,2048,0.006578666468461354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,1536,0.06953599717881945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,16384,0.015498666299713982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,7168,0.02736977736155192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,4096,0.028061333629820082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,7168,0.01037688884470198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,5120,0.012454222473833295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,3072,0.007630222373538547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,16384,0.036074668169021606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,3584,0.014074666632546319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,12288,0.1257893376880222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,1536,0.013255111045307584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,256,0.005952888892756567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,128,0.006223111102978389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,1536,0.016525333126386006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,5120,0.02047288914521535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,16384,0.020577778418858845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,10240,0.04284088810284933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,65536,0.08267644378874037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,2048,0.12237600485483806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,6144,0.012343111137549082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,2560,0.016156444946924847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,7168,0.1580337815814548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,2048,0.006590222318967183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,8192,0.03886399997605218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,4096,0.0469377769364251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,2560,1.1533066431681316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,2560,0.014667555689811707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,65536,0.2195004489686754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,3584,0.02159022291501363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,768,0.018904000520706177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,32,0.0021333334346612296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,2560,0.045498665836122304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,1536,0.010668444136778513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,4096,0.011332444846630096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,3072,0.00793599999613232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,2048,0.013409777647919126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,7168,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,3584,0.010318222145239512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,10240,0.02641422218746609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,768,0.002521777732504739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,3072,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,768,0.03568444318241543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,16384,0.03844088978237576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,2048,0.019927110936906602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,512,0.006057777752478917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,128,0.010438222024175854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,4096,0.007598222129874759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,256,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,64,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,5120,0.025960889127519395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,768,0.0031928889867332247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,64,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,64,0.0035013332962989807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,65536,0.09045866462919448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,12288,0.02269688910908169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,768,0.057576000690460205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,2560,0.007312888900438945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,2048,0.006874666859706243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,8192,0.012823999755912356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,5120,0.024359110328886244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,768,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,4096,0.04883289006021288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,256,0.01614488826857673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,1536,0.006620444357395172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,5120,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,768,0.02163644466135237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,8192,0.046406222714318164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,2048,0.004848888764778773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,12288,0.03943644298447503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,5120,0.009014222357008193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,4096,0.019655111763212416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,8192,0.23413509792751738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,1536,0.006269333263238271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,128,0.004211555752489301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,12288,0.1942462258868747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,1024,0.013755555782053204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,3584,0.007623111208279927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,8192,0.030770666069454614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,12288,0.016537777251667447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,1024,0.01550222271018558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,7168,0.010315555665228102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,5120,0.023583110835817125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,1536,0.015484443969196744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,65536,0.4221626652611627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,128,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,2560,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,4096,0.0041964443193541635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,6144,0.005234666582610872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,2048,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,4096,0.008370666868156856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,64,0.0019324444648292328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,64,0.007628444168302748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,5120,0.018941332896550495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,512,0.0203493336836497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,768,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,65536,0.25460176997714573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,5120,0.12473689185248481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,64,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,4096,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,6144,0.012879110872745514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,64,0.0019066666977273093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,256,0.016209777858522203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,2048,0.0069155556460221606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,512,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,4096,0.011651555697123209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,8192,0.03659910957018534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,512,0.03609777821434869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,1024,0.010203555226325989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,10240,0.036590221855375506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,2048,0.03252977795071072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,128,0.005620444400442972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,10240,0.021632888250880774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,6144,0.033025778002209134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,768,0.030799110730489094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,7168,0.01686311099264357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,2560,0.016736888223224215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,7168,0.009352888498041365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,10240,0.021282666259341772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,7168,0.02163200080394745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,1024,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,512,0.013462222284740873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,7168,0.025246222813924152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,3072,0.33598844210306805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,768,0.011725333001878528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,3584,0.009696000152164036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,32,0.0032435556252797446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,4096,0.0159262220064799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,3584,0.033927109506395124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,5120,0.008775999976529015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,4096,0.017854221993022494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,16384,0.02382755610677931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,12288,0.2638062106238471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,12288,0.0206986665725708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,1024,0.005963555640644497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,256,0.0021724444296624926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,256,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,768,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,1536,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,6144,0.009360888765917884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,7168,0.026379555463790894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,3072,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,5120,0.008296888735559251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,10240,0.029311110575993855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,1024,0.0155102229780621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,16384,0.06565866867701213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,12288,0.014136888914638095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,512,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,4096,0.027963555521435205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,10240,0.05464799867735969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,65536,0.08213066392474704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,12288,0.022164444128672283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,4096,0.00423200014564726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,256,0.011062222222487131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,64,0.005523555394675996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,6144,0.012543111211723752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,64,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,65536,0.06275377670923869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,4096,0.007919111185603673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,16384,0.03614666726854112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,4096,0.007670222057236566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,3584,0.05112888746791416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,768,0.019194665882322524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,256,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,12288,0.3323235511779785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,10240,0.009674666656388177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,10240,0.015784000356992085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,32,0.0052933332820733385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,4096,0.020752888586786058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,128,0.022284444835450914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,768,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,10240,0.04040266738997565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,2560,0.022657778528001573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,1024,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,256,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,6144,0.00906755526860555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,512,0.006401777681377199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,32,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,8192,0.041314668125576444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,6144,0.33459999826219344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,256,0.0147733340660731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,2048,0.031650665733549334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,768,0.00612088872326745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,12288,0.03307111064592997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,6144,0.44838666915893555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,3584,0.007962666451931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,12288,0.09122133255004883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,2560,0.006974221931563483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,3584,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,2560,0.29375023312038845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,768,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,2048,0.05497689048449198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,128,0.028611554039849177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,12288,0.0074026667409473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,1536,0.01477688882086012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,7168,0.012067555553383298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,1024,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,4096,0.007970666719807519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,4096,0.008459555606047312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,1536,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,256,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,128,0.056998219754960805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,6144,0.02374222212367588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,768,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,2560,0.025174222058720056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,1024,0.012523555921183692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,768,0.037145776881111994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,4096,0.021106666988796655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,10240,0.035092443227767944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,1024,0.006221333311663733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,12288,0.33506398730807835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,512,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,6144,0.017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,16384,0.03806311223242018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,7168,0.019296000401178997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,256,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,1024,0.024313777685165405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,2048,0.020581333173645865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,7168,0.018216000662909616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,65536,0.07502222061157227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,768,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,2048,0.026706667410002813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,6144,0.034938666555616595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,768,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,256,0.005554666535721884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,1536,0.022566222482257422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,3584,0.102966222498152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,3584,0.008043555749787224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,12288,0.05334755447175768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,5120,0.016527111331621807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,7168,0.03516799873775906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,32,0.19381422466701934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,256,0.0018622221218215094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,10240,0.31269510587056476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,10240,0.047876444127824574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,128,0.005286222116814719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,512,0.005283555636803309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,3584,0.014857777290874057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,256,0.00350577798154619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,2048,0.014527110589875115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,5120,0.0052675555149714155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,3584,0.06148444281684028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,7168,0.018241778016090393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,256,0.023660444551044043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,1536,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,3072,0.01369955556260215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,5120,0.053787555959489614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,64,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,3584,0.0232622226079305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,512,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,12288,0.013345777988433838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,2560,0.4435599909888373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,128,0.005111111121045219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,512,0.012048000262843238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,10240,0.12082489331563313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,5120,0.04580444428655836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,12288,0.036357333262761436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,32,0.025399999486075506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,128,0.0021075554605987337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,3072,0.07122933202319674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,768,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,7168,0.04000355468855964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,768,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,3584,0.007962666451931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,1536,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,4096,0.01687466601530711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,3072,0.004784000002675586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,10240,0.015871110889646742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,8192,0.00793511089351442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,3072,0.015934222274356417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,12288,0.03695822093221877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,5120,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,12288,0.017869333426157635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,6144,0.0285502208603753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,10240,0.17803555064731172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,128,0.01551822159025404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,2560,0.007272000114123027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,1024,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,256,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,512,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,3584,0.023008000519540574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,3072,0.0075937774446275495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,768,0.039093332158194646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,1024,0.0587680008676317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,1536,0.010116444693671333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,16384,0.06388266881306966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,64,0.005249777601824866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,768,0.018547554810841877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,1024,0.005967111223273807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,5120,0.038482666015625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,512,0.0027671110712819626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,12288,0.06053777535756429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,7168,0.05270488725768196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,16384,0.016148444679048326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,3072,0.0210524449745814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,16384,0.04464088877042135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,4096,0.004217777815130022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,4096,0.08260089159011841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,3584,0.015455111861228943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,6144,0.016103999482260812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,7168,0.009690666364298927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,1024,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,256,0.002132444539003902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,65536,1.2026906543307834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,512,0.041991998751958214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,12288,0.03107200066248576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,64,0.009363555245929295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,2560,0.019478221734364826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,2560,0.007134222322040134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,12288,0.019552888141738046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,6144,0.02224622170130412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,2560,0.007252444823582967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,3584,0.011010666688283285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,32,0.010025777750545079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,512,0.00793511089351442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,1536,0.03768710957633124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,5120,0.015115555789735583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,5120,0.017642666896184284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,10240,0.041249778535630964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,1024,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,65536,0.3483591079711914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,1536,0.017503110898865592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,512,0.020103999310069613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,2048,0.0123848890264829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,32,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,64,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,1536,0.01920088960064782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,1024,0.026368000441127356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,128,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,12288,0.06013422542148166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,7168,0.015846222639083862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,3584,0.017887110511461895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,128,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,3072,0.010341333018408881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,16384,0.057314667436811656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,10240,0.0118231111102634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,128,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,5120,0.03254310952292548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,4096,0.018202667435010273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,6144,0.017536888519922893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,3584,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,7168,0.03809955716133118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,2560,0.00795377790927887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,65536,0.17782933182186553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,3072,0.020647111866209242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,512,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,3072,0.01702222228050232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,4096,0.02404444416364034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,6144,0.023059555225902136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,1024,0.017477333545684814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,1024,0.007659555309348636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,12288,0.41924089855617946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,10240,0.8715733422173394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,1024,0.025584888127115037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,6144,0.04225066635343763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,1024,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,6144,0.020133333073721994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,3072,0.013119999733236102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,128,0.015462223026487561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,5120,0.013567999833159976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,3584,0.028802666399214003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,7168,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,1536,0.008847999903890822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,10240,0.07072089115778606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,8192,0.09709333048926459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,65536,0.07146310806274414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,4096,0.007979555262459649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,1536,0.00850933293501536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,4096,0.020064888728989493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,6144,0.010080888867378235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,65536,0.06416889031728108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,12288,0.10629510879516602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,1024,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,768,0.007670222057236566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,512,0.01925511161486308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,4096,0.017529777354664274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,2048,0.01479111115137736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,3072,0.011375111010339526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,10240,0.022184888521830242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,8192,0.029905776182810467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,768,0.04192177785767449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,6144,0.030080000559488933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,128,0.014471999473041959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,128,0.0025111111915773815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,256,0.014119111829333834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,3072,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,65536,0.11838666598002116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,4096,0.011023110813564725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,8192,0.02808444367514716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,6144,0.006268444574541516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,5120,0.016494222813182406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,1024,0.056977775361802846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,8192,0.013434666726324292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,2048,0.008354666332403818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,12288,0.01850133306450314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,256,0.0038417776425679526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,8192,0.028918223248587713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,4096,0.012434666355450949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,6144,0.02796799937884013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,2560,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,2560,0.013034666577974955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,8192,0.014076444837782117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,12288,0.024360888534122046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,512,0.010421333213647207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,768,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,6144,0.009303111169073317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,7168,0.016864000095261466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,128,0.002987555538614591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,16384,0.7387679947747124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,768,0.00498488876554701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,4096,0.017849778135617573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,2560,0.02458488941192627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,768,0.003850666599141227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,16384,0.03359377715322707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,3584,0.01822933389080895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,256,0.01815111107296414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,256,0.0035377778112888336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,256,0.015861334072219003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,32,0.0019226666125986313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,2048,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,3584,0.04251555601755778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,65536,0.6531350877549913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,7168,0.1749502287970649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,2560,0.008565333154466417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,8192,0.03418044580353631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,5120,0.0655662218729655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,64,0.00216977774269051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,2560,0.03498399919933743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,512,0.007288888924651676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,12288,0.036009778579076133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,512,0.028881778319676716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,10240,0.16789866818322075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,3584,0.007602666815121968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,16384,0.017620444297790527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,768,0.003181333343187968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,256,0.0052728888889153796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,768,0.015139554937680563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,65536,0.28287378946940106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,1536,0.01053955571519004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,512,0.006623111251327727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,65536,0.13465600543551975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,8192,0.7516622013515897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,12288,0.3373342090182834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,8192,0.030088888274298772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,65536,0.06432088878419664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,12288,0.018716444571812946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,3072,0.01094133324093289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,4096,0.012764444781674279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,16384,0.019223110543357003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,10240,0.014039110806253223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,3584,0.008332444561852349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,16384,0.0743288861380683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,4096,0.017873777283562552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,2560,0.01889955500761668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,4096,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,64,0.0058000000814596815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,7168,0.004946666873163647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,16384,0.051368001434538096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,1024,0.00282577777074443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,6144,0.013034666577974955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,1536,0.00666222224632899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,7168,0.011553777588738335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,1024,0.014496000276671516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,1536,0.04961599906285604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,128,0.015184000134468079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,6144,0.03537777728504605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,64,0.0021768889079491296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,65536,0.07269066572189331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,16384,0.13614043924543592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,512,0.005283555636803309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,256,0.02741333345572154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,7168,0.009680888719028896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,7168,0.012748444245921241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,65536,0.4798409144083659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,768,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,256,0.006291555447710886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,2560,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,1536,0.009880888793203566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,4096,0.03525688913133409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,1024,0.0630577802658081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,2048,0.018868444694413077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,1024,0.02197511163022783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,3072,0.008593777815500895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,8192,0.014704888065656027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,2560,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,3072,0.009524444739023844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,4096,0.004524444540341695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,256,0.01479555500878228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,128,0.006291555447710886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,512,0.0028133332315418455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,7168,0.012826667063766055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,16384,0.012759110993809171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,128,0.005335111171007156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,16384,0.24406043688456217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,5120,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,8192,0.011409777734014722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,8192,0.5526000128851997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,128,0.015981333123313058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,1024,0.006035555568006303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,512,0.0038542221817705366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,5120,0.017515555024147034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,12288,0.5598026911417643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,2560,0.00828888929552502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,2560,0.012476444244384766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,1536,0.004744889007674323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,10240,0.07174044185214572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,64,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,4096,0.00490399988161193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,4096,0.008589333130253686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,65536,0.14835111300150552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,6144,0.011438222395049201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,128,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,512,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,4096,0.014198222094111972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,7168,0.038872000243928694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,256,0.007538666327794393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,1536,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,1024,0.006931555353932911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,256,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,1536,0.0062871111763848206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,8192,0.021273778544531927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,256,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,65536,0.07419555717044406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,4096,0.0188391109307607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,32,0.006221333311663733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,5120,0.004885333279768626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,4096,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,65536,0.06896089182959662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,2560,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,1024,0.004520888957712385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,256,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,2560,0.014145778285132514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,6144,0.021943999661339655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,65536,0.07731466823154025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,12288,0.029105779197480943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,3072,0.06220799684524536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,65536,0.08542311191558838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,512,0.0069955554273393415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,4096,0.1081848873032464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,6144,0.0861448910501268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,768,0.0028293333533737394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,7168,0.009975999593734741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,2048,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,512,0.002543111021320025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,16384,0.01614755557643043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,1536,0.01585422290696038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,256,0.002152888932161861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,12288,0.04930311110284594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,2048,0.021218667427698772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,12288,0.01943288909064399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,128,0.006314666734801398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,6144,0.029448890023761328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,3584,0.008441777692900764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,256,0.005401777724424998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,12288,0.041859553919898145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,5120,0.04371822211477491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,256,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,7168,0.015832889411184523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,16384,0.04560977882809109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,4096,0.008637333081828224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,12288,0.013674666484196981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,3584,0.01496444476975335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,65536,4.058229234483507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,768,0.033909334076775446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,768,0.02112355497148302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,128,0.0052888890107472735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,2048,0.008568888737095727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,10240,0.018176888426144917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,512,0.00980000032318963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,6144,0.1692648861143324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,512,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,10240,0.028063111835055884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,10240,0.01588444411754608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,12288,0.03874222106403775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,128,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,10240,0.02568000058333079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,512,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,1024,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,256,0.005547555370463266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,65536,0.2691333293914795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,10240,0.018178666631380718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,3072,0.01867733399073283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,3584,0.012223111258612739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,1536,0.00656355544924736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,12288,0.04496444596184624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,64,0.0018399999373488957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,5120,0.018156444032986958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,768,0.00563466673096021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,16384,0.019697777099079557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,6144,0.03272622161441379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,6144,0.012478222449620565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,128,0.0024888888001441956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,1536,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,6144,0.01260444439119763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,1024,0.007292444507280986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,10240,0.008730666504965888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,512,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,65536,0.0069520001610120135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,768,0.006310222049554189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,12288,0.02199999988079071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,128,0.01924622224436866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,128,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,3584,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,3584,0.010327111515733931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,4096,0.01794577803876665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,3584,0.007602666815121968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,32,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,2560,0.032984889215893216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,256,0.0021768889079491296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,768,0.00793333351612091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,64,0.0018373332503769135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,16384,0.05740355783038669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,768,0.007605333295133378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,128,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,256,0.0022906667242447534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,6144,0.010319111247857412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,16384,0.043291555510626904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,128,0.006622222148709827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,4096,0.020640000700950623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,65536,0.20464444160461426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,1536,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,16384,0.02390577726893955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,512,0.026312889324294195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,2560,0.007296889192528195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,12288,0.029779556724760268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,12288,0.04275733232498169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,5120,0.024529778295093115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,1536,0.1722257799572415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,12288,0.013440888788965015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,10240,0.01643288963370853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,7168,0.034224887688954674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,4096,0.011712888876597086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,3584,0.035593777894973755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,1024,0.009703999592198266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,64,0.005516444643338521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,5120,0.011804444094498953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,10240,0.16309421592288548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,16384,0.023738667368888855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,4096,0.02369955513212416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,7168,0.043619556559456714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,6144,0.012568888564904531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,3072,0.06182666619618734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,8192,0.021686222818162706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,768,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,6144,0.02197244432237413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,1024,0.03401333424780104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,12288,0.02975289026896159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,12288,0.5446809132893881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,2048,0.023354666100607976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,7168,0.010679999987284342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,768,0.015867556134859722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,64,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,6144,0.052355554368760854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,10240,0.02788977821667989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,7168,0.009737778041097853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,64,0.0032960000551409195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,65536,0.08186844322416517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,12288,0.04294755392604404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,1536,0.31792089674207896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,8192,0.04184533490075005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,1024,0.003165333428316646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,5120,0.019259555472267997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,768,0.01802399920092689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,7168,0.016935999194780987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,2048,0.014831999937693277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,2560,0.1204942199918959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,256,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,512,0.012902222573757172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,10240,0.015758223003811307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,12288,0.013463111387358772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,512,0.005968889014588461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,5120,0.01107377807299296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,12288,0.061194664902157254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,3584,0.023074666659037273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,6144,0.012801777985360889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,1024,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,12288,0.013444444371594323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,8192,0.011061333119869232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,2560,0.01681600014368693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,5120,0.013389333254761167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,1536,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,3584,0.034832000732421875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,7168,0.010025777750545079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,16384,0.11576799551645915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,256,0.026271111435360376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,5120,0.07535377475950453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,3584,0.013790222505728403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,8192,0.012963555753231049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,7168,0.013222222526868185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,1024,0.017838221457269456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,4096,0.01443733274936676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,64,0.005613333235184352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,3072,0.0369395547442966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,4096,0.023970666858885024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,64,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,5120,0.013336000343163809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,1536,0.0063288890653186375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,1536,0.005712888720962737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,768,0.00601244428091579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,4096,0.018259555101394653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,768,0.026923555466863845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,3584,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,7168,0.009832888841629028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,6144,0.023992889457278784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,3584,0.011114666859308878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,4096,0.005995555470387141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,4096,0.008008889026112026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,2560,0.008943110704421997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,1024,0.016537777251667447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,2560,0.9246293173895942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,2048,0.017864000466134813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,3072,0.009667555491129557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,1024,0.020618667205174763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,3072,0.0155102229780621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,6144,0.03884177737765842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,12288,0.11336444483862983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,2048,0.025761778155962627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,8192,0.029708445072174072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,6144,0.03007733490731981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,2048,0.009921777579519484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,5120,0.022848889231681824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,64,0.0018177777528762817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,7168,0.24264889293246797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,65536,0.07597600089179145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,16384,0.042562666866514415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,7168,0.011033777561452655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,2560,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,1024,0.004211555752489301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,65536,0.10096000300513373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,12288,0.018214222457673814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,256,0.0021706666383478376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,4096,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,256,0.021752888957659405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,12288,0.030761778354644775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,4096,0.0199937770764033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,512,0.003506666670242945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,65536,0.9749093585544162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,3584,0.010703110860453712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,3072,0.012770666844315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,4096,0.05189333359400431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,2560,0.013437333206335703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,64,0.023411555422676936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,4096,0.014283556077215405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,7168,0.010243555737866296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,128,0.0155102229780621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,5120,0.03910400138960944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,64,0.001869333287080129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,4096,0.020209777686331008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,3584,0.014188443620999655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,5120,0.01689511040846507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,3584,0.010863111250930362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,16384,0.03283111254374186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,7168,0.013464888764752282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,3072,0.008007110820876228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,1536,0.009694221946928237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,512,0.020239111449983384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,16384,0.04790488878885905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,1024,0.003173333282272021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,12288,0.08887377712461685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,10240,0.01684711045689053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,7168,0.008106666306654612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,3584,0.019160888261265226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,64,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,16384,0.07053511010275947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,512,0.005869333528810077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,3584,0.03758488761054145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,16384,0.03748800026045905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,3584,0.015169777803950839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,128,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,512,0.011749332977665795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,2048,0.006984000404675801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,768,0.01758311192194621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,4096,0.010614222122563256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,3072,0.007974222302436829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,128,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,5120,0.016164445214801364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,12288,0.36639822853936094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,5120,0.012748444245921241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,5120,0.01072177787621816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,16384,0.02125511070092519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,65536,10.69903564453125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,3072,0.021225778592957392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,65536,0.03327555457750956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,2048,0.009975999593734741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,2560,0.16006844573550755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,1536,0.014789332946141561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,10240,0.07239288753933378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,64,0.008159110943476358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,1536,0.009777777724795872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,768,0.0028391112056043413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,4096,0.0259253333012263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,2048,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,768,0.003526222374704149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,768,0.02994133366478814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,768,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,512,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,768,0.006039111150635614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,256,0.005364444520738389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,2560,0.013433777623706393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,2560,0.027665777338875666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,65536,0.3122319910261366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,3584,0.3940257761213515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,512,0.006766222417354584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,2048,0.014381332529915703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,4096,0.009813333551088968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,5120,0.011400889191362592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,7168,0.03747999999258254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,10240,0.039657778210110135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,16384,0.022865777214368183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,3584,0.009096889032257928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,16384,0.01856355534659492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,64,0.004523555437723796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,12288,0.12857245074378118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,8192,0.013841778039932251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,16384,0.04502044452561272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,2560,0.006973333656787872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,16384,0.028816888729731243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,64,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,3072,0.030551112360424463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,32,0.0025271111064487034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,2048,0.005201777650250329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,7168,0.049591110812293164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,8192,0.02234222160445319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,6144,0.020768000019921195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,2048,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,768,0.01851288808716668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,5120,0.008564444051848518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,2048,0.0063999998900625445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,3584,0.016921778519948322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,3072,0.07289333475960626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,65536,0.057746668656667076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,4096,0.011870221959220039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,5120,0.10617955525716145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,10240,0.028071112102932397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,65536,0.22920534345838758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,10240,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,3584,0.035840001371171736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,7168,0.010321777727868822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,256,0.014471999473041959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,128,0.005180444568395615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,16384,0.016044444508022733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,3072,0.02861422300338745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,8192,0.023563555545277063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,256,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,32,0.0025351111673646504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,2560,0.016492444607946608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,12288,0.0630542238553365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,5120,0.01961688862906562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,7168,0.01773244473669264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,12288,0.23808622360229492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,2048,0.01239733315176434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,128,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,128,0.18159111340840658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,3584,0.014394667413499622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,128,0.015344889627562629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,2560,0.02534666657447815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,6144,0.014150222142537435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,256,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,32,0.007624000310897827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,1536,0.009907555249002244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,12288,0.03117777903874715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,128,0.0121760004096561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,256,0.004895110925038655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,512,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,256,0.014127110441525778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,16384,0.020988444487253826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,4096,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,10240,0.02989333205752903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,3584,0.031599998474121094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,7168,0.01033333357837465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,1536,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,5120,0.01680266691578759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,10240,0.01648177703221639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,10240,0.14168266455332437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,1024,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,256,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,2048,0.013429332938459186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,1536,0.018216888109842937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,3072,0.007598222129874759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,10240,0.01759022143152025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,2048,0.013083555632167392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,12288,0.034825778669781156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,768,0.006539555473460092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,2048,0.02030755579471588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,768,0.00351555562681622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,65536,0.13775555292765299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,2560,0.007251555720965068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,4096,0.04826755656136406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,2048,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,8192,0.04360266526540121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,3072,0.017107556263605755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,6144,0.025758221745491028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,768,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,8192,0.21766044033898246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,65536,2.4114649030897355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,3072,0.019578667150603402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,6144,0.012418666647540199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,65536,0.06308088699976604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,5120,0.021615111165576514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,6144,0.012034667034943899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,65536,0.06754755311542086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,1024,0.003216888962520493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,3584,0.10690666569603814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,32,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,5120,0.014099554883109199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,2560,0.01666755477587382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,512,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,2560,0.028824888997607764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,3072,0.044369777043660484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,16384,0.4417733351389567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,3072,0.007557333343558841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,1536,0.01648177703221639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,65536,0.09196444352467854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,8192,0.011428444749779172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,65536,0.42866844601101345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,7168,0.01420177850458357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,3072,0.038450668255488075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,128,0.014848889576064216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,64,0.0028275555620590844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,3584,0.007583111524581909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,12288,0.004931555440028508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,10240,0.0277839998404185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,3072,0.02566755480236477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,16384,0.01216533366176817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,3584,0.014287999934620328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,2560,0.034028444025251604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,1024,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,6144,0.014069333672523499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,2560,0.009472889204819998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,8192,0.03564800156487359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,512,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,4096,0.13819200462765163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,16384,0.018033777674039204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,32,0.005748444547255834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,10240,0.09550044271681045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,256,0.005636444522274866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,3584,0.14157599873012966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,2048,0.02775466607676612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,5120,0.006718222051858902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,1024,0.02168088820245531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,3584,0.008661333057615492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,512,0.22069511148664686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,1536,0.02234488891230689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,3584,0.08235822121302287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,16384,0.027286221583684284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,2560,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,3072,0.007972444097201029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,2560,0.05682933330535889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,5120,0.07492355505625407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,10240,0.012320888539155325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,256,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,128,0.006688000013430913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,3584,0.00832622249921163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,7168,0.015511999527613321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,3584,0.009346666435400644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,2560,0.009297777381208207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,6144,0.004816888935036129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,512,0.0058764442801475525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,12288,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,10240,0.0255422227912479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,65536,0.03443377878930833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,5120,0.01681155628628201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,8192,0.052914665804968945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,4096,0.005980444451173146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,7168,0.016155555844306946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,6144,0.026445334156354267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,2048,0.006591111007663939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,2048,0.016134222348531086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,2048,0.01239111108912362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,12288,0.013763555222087435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,65536,0.9222382439507378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,5120,0.043599112166298755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,10240,0.08697155449125503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,5120,0.026464000344276428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,65536,0.39306312137179905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,65536,0.2408693366580539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,1536,0.04483111037148369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,1536,0.016757332616382174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,6144,0.20996177196502686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,65536,0.5102791256374782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,256,0.005411555369695027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,5120,0.01921777758333418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,2560,0.02221155497762892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,1024,0.007947555846638149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,7168,0.02091111077202691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,3072,0.007728888756699032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,3072,0.007882666256692674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,10240,0.030360000001059637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,5120,0.04455822375085619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,2560,0.02509422269132402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,8192,0.030375109778510198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,2560,0.014037332601017423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,5120,0.08996799919340347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,2048,0.12980533970726862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,4096,0.022832888695928786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,256,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,6144,0.023038221730126276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,6144,0.03666222095489502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,12288,0.011711999773979187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,12288,0.013766222529941134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,128,0.01512888901763492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,6144,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,128,0.00453333349691497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,3072,0.029429333077536687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,3072,0.08481155501471625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,10240,0.0439591109752655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,512,0.025245333711306255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,4096,0.013656888571050433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,128,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,7168,0.017536888519922893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,1536,0.016532444291644625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,12288,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,5120,0.06733866532643636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,64,0.0021928888228204516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,6144,0.49542490641276044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,8192,0.022612444228596155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,3584,0.013168888787428537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,10240,0.21320711241828072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,8192,0.036658667855792575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,3584,0.012755555411179861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,65536,0.0853875544336107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,1536,0.006551110910044775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,3584,0.016135111451148987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,5120,0.024772443705134924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,7168,0.04127111037572225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,7168,0.018617777360810172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,8192,0.011013333168294696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,1536,0.17081956068674722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,2048,0.016400888562202454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,16384,0.05910755528344048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,8192,0.005762666463851929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,6144,0.014241778188281588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,3072,0.011697777443461947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,768,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,10240,0.008982222113344405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,12288,0.02824977702564663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,6144,0.08660088645087348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,2048,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,3072,0.007246221933099959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,4096,0.03968977928161621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,64,0.005789333333571752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,768,0.07036888599395752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,5120,0.021937777598698933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,10240,0.004912000149488449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,32,0.0030008889734745026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,2048,0.01648266613483429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,768,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,512,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,8192,0.018416888184017606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,256,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,1536,0.03613955444759793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,3584,0.007958222594526079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,10240,0.09692977534400092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,4096,0.015423110789722867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,256,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,768,0.019882667395803664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,3072,0.016513778103722464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,10240,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,7168,0.0890008873409695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,256,0.015987555185953777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,7168,0.006728000111050076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,4096,0.01793155570824941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,8192,0.09060355689790513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,8192,0.035346666971842446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,65536,0.009341333475377824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,16384,0.2747795581817627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,3072,0.007888000044557784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,32,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,768,0.007239999870459239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,12288,0.07198755608664618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,3072,0.020917332834667627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,3584,0.007807110746701558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,10240,0.023189332750108507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,12288,0.039974222580591835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,3072,0.13413688871595594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,2560,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,2048,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,128,0.005584888988071018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,2048,0.012390221986505719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,2560,0.010325333310498131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,7168,0.02642044425010681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,10240,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,5120,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,12288,0.01420177850458357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,10240,0.012274666792816587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,8192,0.0351200004418691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,65536,0.6260799831814235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,64,0.0018524444765514797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,3584,0.00833777752187517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,4096,0.013401777380042605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,12288,0.0191600008143319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,64,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,10240,0.017318222257826064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,3584,0.01902577777703603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,1024,0.0028702221396896574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,4096,0.024392000503010217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,256,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,6144,0.027007111244731482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,16384,0.017997332745128207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,3584,0.010085333552625444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,6144,0.028071112102932397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,4096,0.00812000036239624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,5120,0.00833777752187517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,10240,0.01573511130279965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,6144,0.1333928902943929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,3584,0.019834667444229126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,16384,0.05786755349900988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,64,0.0020711111525694528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,8192,0.009359111388524374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,3584,0.10239377948972915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,2048,0.004652444273233414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,4096,0.006620444357395172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,128,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,32,0.0103502223889033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,2560,0.10606133275561863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,32,0.0018044445249769424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,65536,0.010616000327799054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,2048,0.06934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,32,0.01036000003417333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,12288,0.013413333230548434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,32,0.0024604443460702896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,768,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,6144,0.026132444540659588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,1536,0.017176000608338248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,5120,0.01035199976629681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,65536,0.09393688705232407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,1024,0.015471110741297403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,128,0.008281778130266402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,5120,1.9763724009195964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,6144,0.01701866587003072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,12288,0.0366613335079617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,4096,0.006948444578382704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,7168,0.013250666360060373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,1536,0.009006222089131674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,32,0.025998221503363714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,1024,0.009466666314336989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,1536,0.008664000365469191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,512,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,6144,0.019983111156357657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,256,0.0069715554515520734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,5120,0.012958221965365939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,12288,0.02010222276051839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,2560,0.02737599942419264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,12288,0.017857778403494094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,4096,0.00831288927131229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,1024,0.018233777748213876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,12288,0.01812177730931176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,16384,0.01664533383316464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,16384,0.0359919998380873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,3584,0.007329777710967594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,32,0.001820444439848264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,3584,0.04198044538497925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,256,0.007904888855086433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,768,0.010935111178292168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,8192,0.013420444395807056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,3072,0.0074986666440963745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,768,0.02202577723397149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,12288,0.27019111315409344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,8192,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,128,0.0021742222209771476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,12288,0.34136001269022626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,512,0.013595555391576556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,1024,0.02278222143650055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,3072,0.014485332700941296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,7168,0.010327999790509542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,16384,0.11578667163848877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,16384,0.011022222538789114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,768,0.009416888985368941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,3584,0.022227555513381958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,768,0.003863999827040566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,65536,0.06606400012969971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,2048,0.0203128887547387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,256,0.023711111810472276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,2048,0.035367111365000405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,7168,0.011015111373530494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,4096,0.011979555918110741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,256,0.015276443627145557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,5120,0.023338667220539514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,10240,0.051162666744656034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,10240,0.7316453191969129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,6144,0.03273866573969523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,3072,0.007734222544564142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,12288,0.007287110719415877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,64,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,512,0.00685511115524504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,512,0.008975110948085785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,8192,0.07683110899395414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,64,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,128,0.022776888476477727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,65536,0.029329776763916016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,1024,0.009186666872766282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,10240,0.01220266686545478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,1536,0.022444445225927565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,5120,0.049010667535993785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,8192,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,768,0.015451555450757345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,128,0.011039999624093374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,2560,0.018997333116001554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,1024,0.006303999986913469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,1536,0.007296000089910295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,3584,0.014919999572965833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,8192,0.06751644611358643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,5120,0.271304898791843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,7168,0.010019555687904358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,5120,0.017097777790493436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,12288,0.033109333780076765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,6144,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,4096,0.008923555413881937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,2048,0.004181333300140169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,7168,0.016711110870043438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,8192,0.011066666907734342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,3584,0.01514844430817498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,768,0.019856888386938307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,1536,0.0069520001610120135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,128,0.014458666245142618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,2560,0.013081777426931592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,5120,0.012053333222866058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,128,0.014807111687130399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,3072,0.026710222164789837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,256,0.007920888562997183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,8192,0.01826844447188907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,3584,0.03310133351220025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,16384,0.009309333231714036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,16384,0.05857511361440023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,7168,0.040545778142081365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,8192,0.050360000795788236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,12288,0.019654222660594516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,32,0.0024844445288181305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,2560,0.013213333156373767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,3584,0.024872889121373493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,8192,0.010453333457310995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,6144,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,768,0.018173333671357896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,512,0.005394666559166378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,3584,0.015450666348139444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,10240,0.2365137736002604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,6144,0.04217333263821072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,2560,0.004874666531880696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,16384,0.05032889048258463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,10240,0.004858666823969947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,3072,0.08189333147472806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,16384,0.04114311271243625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,1536,0.01200355589389801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,768,0.04314400090111626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,65536,0.19726399580637613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,8192,0.06814844740761651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,1536,0.028483556376563177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,1536,0.006455111006895701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,2048,0.013065777719020844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,1536,0.09530844291051228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,4096,0.028115553988350764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,65536,0.10200355450312297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,16384,0.009899555808968013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,16384,0.0474639998541938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,1536,0.09899644719229804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,7168,0.07370133532418145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,64,0.00796888851457172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,2048,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,4096,0.00942666663063897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,3072,0.004932444542646408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,1024,0.006222222414281633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,12288,0.027084444959958393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,5120,0.028083556228213843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,512,0.005423111220200856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,7168,0.06601866748597887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,256,0.0031537777847713898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,65536,0.07599911424848768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,4096,0.006053333481152852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,5120,0.008274666965007782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,5120,0.01517599986659156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,6144,0.06972622209125094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,10240,0.0624106658829583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,1024,0.015005333556069268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,3072,0.012642666697502136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,128,0.002130666747689247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,2560,0.02161066730817159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,6144,0.1772986650466919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,10240,1.1977092954847548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,2560,0.017532444662517972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,1024,0.027504000398847792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,32,0.007252444823582967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,32,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,2048,0.006523555351628198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,8192,0.18401155206892225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,8192,0.015631111131774057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,6144,0.016523554921150208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,256,0.016939555605252583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,65536,0.085207998752594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,3072,0.017156443662113614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,7168,0.012286221815480126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,1536,0.007341333561473423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,2048,0.00453333349691497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,10240,0.026360889275868733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,512,0.0021893332401911416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,1536,0.015799999237060547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,65536,0.30273956722683376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,12288,0.13310221831003824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,3072,0.004927111168702443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,4096,0.019311999281247456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,3584,0.00831288927131229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,64,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,256,0.024039111203617517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,2560,0.007377777662542131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,3584,0.006912888752089606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,3584,0.041131556034088135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,16384,0.030834668212466772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,768,0.003527111063400904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,768,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,12288,0.21305599477556017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,16384,0.012078222301271228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,768,0.017599110802014668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,2560,0.007088888850477006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,2048,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,32,0.003176888864901331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,1536,0.017163554827372234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,256,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,256,0.017516444126764934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,10240,0.02092444399992625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,16384,0.026240888569090102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,1024,0.006396444307433234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,16384,0.015829333000712924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,3072,0.031410667631361223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,10240,0.010581333604123859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,8192,0.022117333279715642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,128,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,8192,0.024047111471494038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,10240,0.05856000052558052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,12288,0.01959911154376136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,32,0.005234666582610872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,2560,0.0074346669846110875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,2048,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,2048,0.006992888947327931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,10240,0.018637332651350234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,512,0.014488000008794995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,65536,0.1351084444257948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,16384,0.016519111063745286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,2048,0.006329777754015393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,256,0.016175111134847004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,6144,0.008955555657545725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,1024,0.015951111912727356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,2048,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,7168,0.03179733289612664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,7168,0.39581778314378524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,2560,0.015799999237060547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,5120,0.06178310844633314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,32,0.00630666646692488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,16384,0.0788559979862637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,2048,0.024756444825066462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,8192,0.01686755485004849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,768,0.003536888708670934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,12288,0.028959999481836956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,32,0.009383111364311641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,2048,0.030672887961069744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,128,0.014897776974572076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,8192,0.29230578740437824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,4096,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,3072,0.021687111920780603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,10240,0.03623822331428528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,1024,0.01240888900227017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,128,0.010087110930018956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,4096,0.01995644470055898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,256,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,6144,1.065490616692437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,2048,0.008949333594905006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,1024,0.01310488912794325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,4096,0.010994666980372535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,4096,0.013978666729397245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,512,0.023962666591008503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,5120,0.010023111270533668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,2560,0.007990222010347579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,3072,0.055607113573286265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,3072,0.009154666629102495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,6144,0.024626667300860088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,512,0.02786133355564541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,128,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,128,0.00490044429898262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,4096,0.008645333349704742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,512,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,512,0.015437333120240105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,65536,0.10233689016766018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,6144,0.01440000037352244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,12288,0.10991555452346802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,6144,0.019606222709019978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,16384,0.03976000017589993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,4096,0.008299555215570662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,512,0.017160000072585214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,6144,0.005000000198682149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,128,0.0055440002017551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,8192,0.011639110743999481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,7168,0.012376888758606382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,2048,0.0069848886794514125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,4096,0.00999111102686988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,768,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,65536,0.06458577844831678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,256,0.006622222148709827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,128,0.005944889038801193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,32,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,65536,0.0912604464424981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,2048,0.014543111125628153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,12288,0.029640889830059473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,5120,0.04873777760399712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,8192,0.03822311096721225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,7168,0.027036445008383855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,6144,0.036716444624794856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,6144,0.016142222616407607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,10240,0.01240888900227017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,1536,0.0035075553589397003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,10240,0.03696622120009528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,512,0.00925333301226298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,512,0.005910222315125995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,8192,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,3072,0.007339555356237624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,6144,0.04252800014283922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,128,0.030838221311569214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,3584,0.0075048887067370946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,768,0.004936000125275718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,256,0.013254221942689685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,2560,0.04420266548792521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,256,0.002512000087234709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,768,0.01740800009833442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,65536,0.4956640137566461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,4096,0.03216444452603658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,256,0.00563466673096021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,6144,0.025547555751270715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,12288,0.17006044917636445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,12288,0.01348888874053955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,256,0.022340445054901972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,5120,0.14416621790991888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,64,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,4096,0.07518133189943102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,65536,0.1916924450132582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,7168,0.011035555766688453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,3072,0.06303644180297852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,3072,0.00741511086622874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,256,0.025211556090248957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,1024,0.007074666519959767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,512,0.015106666419241162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,32,0.0019342222561438878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,65536,0.2382240030500624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,1536,0.013539555172125498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,768,0.005437333136796951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,7168,0.004942222187916438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,128,0.002509333400262727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,256,0.003528888854715559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,8192,0.04227377639876472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,12288,0.035036444664001465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,7168,0.032602667808532715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,256,0.02862488892343309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,6144,0.6096728642781576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,16384,0.024175110790464614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,512,0.014520888527234396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,4096,0.01237777786122428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,12288,0.24013243781195748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,5120,0.013632888595263163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,256,0.01513777838812934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,3072,0.011702222128709158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,256,0.005554666535721884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,128,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,32,0.0057751110030545135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,6144,0.012423999607563019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,512,0.00564444437623024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,3584,0.010605333579911126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,5120,0.009638222555319468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,65536,0.5909306738111708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,3584,0.016905777984195285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,2048,0.0069022224181228215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,8192,0.04259999924235874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,8192,0.023634667197863262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,2048,0.007270221908887227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,2048,0.010983111129866706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,2560,0.02574399941497379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,128,0.002838222309947014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,8192,0.06755466593636407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,12288,0.028005331754684448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,1024,0.03071111109521654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,12288,0.013730666703648038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,2560,0.008548444343937768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,16384,0.02871555421087477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,1024,0.004516444272465176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,1024,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,5120,0.008904889225959778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,5120,0.013110222087966071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,2048,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,256,0.01718133356836107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,65536,0.17527555094824898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,1536,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,2048,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,12288,0.01812622282240126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,5120,0.029698666599061754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,512,0.005631111148330901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,256,0.005728888842794631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,3072,0.022343110707071092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,8192,0.023732443650563557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,512,0.008993777963850234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,8192,0.03985777828428481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,65536,0.06974755393134223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,768,0.010968888799349466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,1024,0.006974221931563483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,65536,0.07218488719728258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,16384,0.03668444355328878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,1024,0.0031884445084465873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,1536,0.0035333335399627686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,768,0.0069324444565508105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,128,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,256,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,3584,0.015875554747051664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,1024,0.005629333357016246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,12288,0.05048888921737671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,10240,0.020624000165197585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,4096,0.025429333249727886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,7168,0.019263111882739596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,512,0.0038942222793896994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,12288,0.014123555686738757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,8192,0.03518133362134298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,2560,0.0069324444565508105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,1536,0.0038373333712418876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,1536,0.00351200004418691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,128,0.004024888906213972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,8192,0.016015110744370353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,6144,0.01760444376203749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,5120,0.009000000026490953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,32,0.0021075554605987337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,128,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,1024,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,8192,0.014346667461925082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,768,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,512,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,1536,0.0034924443397257063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,10240,0.1693200005425347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,4096,0.03260711166593764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,5120,0.024420445164044697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,64,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,2560,0.007975111405054728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,10240,0.046698666281170316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,16384,0.024678222007221643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,64,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,512,0.004008888784382078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,6144,0.06994222270117866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,2560,0.008473777936564552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,5120,0.016522667474216886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,1536,0.0038648889296584656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,512,0.007589333587222629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,256,0.012164444559150271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,256,0.061273780133989125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,64,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,16384,0.2821253405676948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,32,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,4096,0.07162844472461276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,256,0.0022177776942650476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,2048,0.026908444033728704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,3072,0.008310221963458592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,5120,0.017068444026841056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,12288,0.028833776712417603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,64,0.0018133332745896445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,768,0.014164444473054675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,768,0.008112000094519721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,6144,0.016304888659053378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,768,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,256,0.019869332512219746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,3584,0.004577777865860197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,12288,0.06072977516386244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,4096,0.00792266676823298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,64,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,256,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,16384,0.17755999830034044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,256,0.012054222325483957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,10240,0.08908266491360134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,32,0.0023271110322740343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,4096,0.28355733553568524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,6144,0.020941333638297185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,128,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,16384,0.11448355515797932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,6144,0.011363555159833698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,3584,0.014533332652515836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,32,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,64,0.0028204443968004654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,7168,0.017331555485725403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,256,0.002516444358560774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,3072,0.034466667307747736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,2048,0.01443377799457974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,6144,0.01089599976936976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,2048,0.0069715554515520734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,8192,0.023399111297395494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,64,0.005429333282841577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,6144,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,12288,0.014429334137174817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,256,0.002130666747689247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,1536,0.03570133447647095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,4096,0.030006223254733618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,8192,0.03985333442687988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,12288,0.02812266680929396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,3584,0.03572355707486471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,16384,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,4096,0.11703288555145264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,65536,0.009014222357008193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,3584,0.00741244438621733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,768,0.030792001220915053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,3584,0.007963555554548899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,2560,0.0052737775776121355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,8192,0.11645421716901992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,2048,0.04952266812324524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,256,0.00499555551343494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,256,0.005872000008821487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,65536,0.22068088584476045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,5120,0.03334666623009576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,7168,0.021581333544519212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,1024,0.021584888299306233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,7168,0.01163200040658315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,12288,0.02401155564520094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,8192,0.01275466630856196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,2048,0.013478221992651621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,5120,0.07710666788948907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,12288,0.0366613335079617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,5120,0.04673244555791219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,512,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,65536,0.8044453197055392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,1536,0.02763644523090786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,512,0.005533333453867171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,16384,0.02141155633661482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,3072,0.019132445255915325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,10240,0.02550933261712392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,768,0.0031715554909573663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,10240,0.10285333130094741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,128,0.007589333587222629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,32,0.00796888851457172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,65536,0.09133333630032009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,2048,0.013167999684810638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,2048,0.030163556337356567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,32,0.007620444728268518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,7168,0.035642666949166193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,2048,0.006319111006127463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,512,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,5120,0.01607644392384423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,5120,0.03290844294759963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,12288,0.028743998871909246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,7168,0.012503999802801343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,8192,0.037456889947255455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,6144,0.009203555683294931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,2560,0.06827111376656426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,2560,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,64,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,16384,0.02163999941613939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,3072,0.035382221142450966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,16384,0.03407999873161316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,4096,0.008632889224423302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,8192,0.00812177773978975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,64,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,12288,0.0480693346924252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,512,0.036363555325402155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,256,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,5120,0.05500088797675239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,8192,0.015467555986510383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,64,0.006614222294754452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,16384,0.013816888961527081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,512,0.01905866629547543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,3584,0.007736889024575551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,10240,0.04088888896836175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,3072,0.01719911065366533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,512,0.034041778908835515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,512,0.006923555499977536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,256,0.015121777852376303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,16384,0.0880240003267924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,16384,0.02164977788925171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,3584,0.012129777835475074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,4096,0.004519111166397731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,1024,0.009671111073758867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,32,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,10240,0.014359111587206522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,64,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,256,0.00965244405799442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,2048,0.017883555756674874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,7168,0.011708444191349877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,8192,0.01312355531586541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,8192,0.02292355563905504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,7168,0.014736889137162102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,10240,0.026721777187453374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,7168,0.0203475554784139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,512,0.005633777628342311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,3584,0.015477332803938123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,128,0.010409778190983666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,2560,0.014333332578341166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,3584,0.017055999901559617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,3072,0.012522666818565793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,512,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,768,0.017242666747834947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,1024,0.007299555672539606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,1024,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,3072,0.0053084443012873335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,5120,0.009328000247478485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,12288,0.12477866808573405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,1536,0.01513777838812934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,1024,0.004022222426202563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,512,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,32,0.005584888988071018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,512,0.006578666468461354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,65536,0.1844151152504815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,256,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,768,0.016903110676341586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,16384,0.04723111126157972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,5120,0.01274577776590983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,128,0.013784888717863293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,3584,0.007612444460391998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,768,0.008117333054542542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,65536,0.07480800151824951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,64,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,3072,0.015825778245925903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,1536,0.01824711097611321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,10240,0.01573422220018175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,8192,0.033944000800450645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,1536,0.003549333247873518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,64,0.002199999988079071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,10240,0.12893244955274794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,1536,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,768,0.030371556679407757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,3584,0.031247112486097548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,12288,0.026781333817376032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,512,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,768,0.017455110947291057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,7168,0.008072888685597314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,65536,0.20703200499216715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,32,0.02440622283352746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,7168,0.026288888520664636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,256,0.0021840000732077495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,12288,0.010088889135254754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,2560,0.026353778110610113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,8192,0.444926208919949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,32,0.004866666677925321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,1536,0.006632888896597757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,3072,0.0674826635254754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,128,0.03741777936617533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,8192,0.739133358001709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,1536,0.014455111490355598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,2048,0.009310222334331935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,3584,0.008188444707128737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,2560,0.0614924430847168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,2560,0.01036711119943195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,10240,0.0181022220187717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,10240,0.17215289009941948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,4096,0.019175110591782462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,5120,0.19536799854702422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,256,0.0021333334346612296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,12288,0.07546666595670912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,2048,0.02056711084312863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,7168,0.026957333087921143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,6144,0.007920888562997183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,7168,0.08127822478612264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,16384,0.012136889000733694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,8192,0.013021333350075616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,12288,0.21252711613972983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,2560,0.008672888908121321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,12288,0.019153777096006606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,3072,0.007785778078767989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,32,0.0018586667461527719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,64,0.012736000120639801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,2048,0.017779555585649278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,5120,0.004603555632962121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,1536,0.006296000132958095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,128,0.006305777778228124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,1024,0.007615110940403408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,3072,0.017419555120997958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,64,0.0018355554590622585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,7168,0.013336000343163809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,16384,0.0738640030225118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,65536,0.14217066764831543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,1536,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,128,0.010093332992659675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,2048,0.023200889428456623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,7168,0.14285422696007624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,2560,0.009992000129487779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,3584,0.1251573297712538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,4096,0.02485688858562046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,65536,0.1194302241007487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,6144,0.009338666995366415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,12288,0.25232354799906415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,32,0.00217333332531982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,256,0.015827554795477126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,128,0.008646222452322641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,2560,0.006950221955776215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,768,0.007631111476156447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,2560,0.013081777426931592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,3072,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,65536,0.08650044600168864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,3584,0.018346667289733887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,3072,0.0165928883685006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,4096,0.019723556107944913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,5120,0.030248890320460003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,128,0.0028124445428450904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,5120,0.03663911090956794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,2560,0.01035644445154402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,32,0.0018568889548381169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,1536,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,10240,0.01898755629857381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,8192,0.012079111403889127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,5120,0.008625778059164682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,3584,0.016530666086408827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,12288,0.017998221847746108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,65536,0.09142222007115682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,128,0.014143110977278816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,32,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,2048,0.011975111232863532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,4096,0.007869333028793335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,256,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,32,0.00180177783800496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,8192,0.08153155777189466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,2048,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,1536,0.01717866626050737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,128,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,10240,0.048347555928760104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,768,0.030415998564826116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,65536,0.06762578090031941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,5120,0.008965333302815756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,6144,0.070997330877516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,1024,0.009668444593747457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,7168,0.024365334047211543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,64,0.0018044445249769424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,32,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,2048,0.005650666852792104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,32,0.007997333175606197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,512,0.04087200098567539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,6144,0.018009778526094224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,5120,0.011362666885058085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,3584,0.035623110002941556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,12288,0.06925688849555121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,4096,0.007815111014578078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,4096,0.00903200027015474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,5120,0.008997333546479544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,8192,0.0155395550860299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,768,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,128,0.0021600000974204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,2048,0.05064800050523546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,65536,2.4337857564290366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,32,0.0068959999415609576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,1536,0.03881244526969062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,4096,0.017843556072976854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,2560,0.015105777316623263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,2560,0.0052693333062860705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,3584,0.007923555870850882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,768,0.007616000043021307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,12288,0.01995644470055898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,12288,0.023383999864260357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,256,0.009597332941161262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,1024,0.015643555257055495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,768,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,128,0.008277333445019191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,16384,1.0496773189968533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,1536,0.02852977646721734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,1024,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,128,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,1024,0.006213333457708359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,256,0.005036444299750858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,7168,0.00832266691658232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,7168,0.024668445189793903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,768,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,256,0.0021191111041439907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,3072,0.29107022285461426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,768,0.013354666531085968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,2048,0.013426666458447775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,2048,0.09974044561386108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,6144,0.01776355504989624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,2048,0.006632888896597757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,2560,0.5711901982625326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,1024,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,128,0.01035199976629681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,5120,0.025850666893853083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,768,0.006235555642180973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,12288,0.038126223617129855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,6144,0.023338667220539514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,12288,0.0351235568523407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,5120,0.1727386713027954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,2048,0.008978666530715095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,6144,0.009302222066455418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,3584,0.02488533324665493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,512,0.0025191110455327565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,3072,0.017408889200952318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,1024,0.011624000138706632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,1024,0.0058764442801475525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,7168,0.04211644331614176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,6144,0.013419555293189155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,16384,0.08305333058039348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,1536,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,7168,0.010041777458455827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,10240,0.08630933364232381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,16384,0.03251288996802436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,3072,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,65536,0.2002506653467814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,1536,0.022351110974947613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,1536,0.006785777707894643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,768,0.00281866660548581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,5120,0.025790222816997107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,5120,0.03158844510714213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,8192,0.010678222609890832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,10240,0.9621253543429904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,768,0.036120888259675764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,4096,0.008721777962313758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,32,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,65536,0.04968800147374471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,1024,0.02777333392037286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,512,0.02700622214211358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,7168,0.31859556833902997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,64,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,7168,0.01015466699997584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,128,0.005943999936183293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,16384,0.05223555697335137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,7168,0.01240888900227017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,32,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,1536,0.0038399998512532976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,32,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,12288,0.13510755697886148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,32,0.007962666451931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,12288,0.061894224749671094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,2560,0.006809777683681912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,128,0.003504000190231535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,4096,0.01576266686121623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,7168,0.17767732673221162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,1536,0.019488889310095046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,768,0.020117332537968952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,2560,0.00703111125363244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,512,0.002495999965402815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,3584,0.2013982269499037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,65536,0.3208204375372993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,768,0.0421653323703342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,2560,0.020747555626763236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,128,0.03243644369973077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,10240,0.01856355534659492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,2560,0.03643111056751675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,16384,0.01980533368057675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,65536,0.14549688498179117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,1536,0.009199110998047723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,12288,0.014144000079896716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,8192,0.05025777882999844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,10240,0.012086222569147745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,12288,0.03530044356981913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,65536,0.1559013393190172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,8192,0.01739377776781718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,2048,0.013790222505728403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,3072,0.007967111137178209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,128,0.006620444357395172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,2048,0.025452444950739544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,65536,0.1212186680899726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,512,0.012891555825869242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,3584,0.008332444561852349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,3072,0.014474666780895658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,8192,0.015290665957662793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,4096,0.008626666333940294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,12288,0.017024000485738117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,512,0.013747555514176687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,12288,0.058487998114691846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,8192,0.010646222366227044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,1024,0.003818666769398583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,128,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,5120,0.01822933389080895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,128,0.03754933343993293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,256,0.005218666460778978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,256,0.02277244461907281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,256,0.006591111007663939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,16384,0.016507556041081745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,65536,0.046894222497940063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,1536,0.015187554889255099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,32,0.017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,32,0.005787555542257097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,16384,0.03145866592725118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,65536,0.1751075585683187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,512,0.012134222520722283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,128,0.014288889037238227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,6144,0.03459377752410041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,8192,0.016454221473799813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,256,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,1536,0.006243555496136348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,6144,0.04984800020853678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,768,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,128,0.011687999798191918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,12288,0.4249555534786648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,5120,0.0042568888101312844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,1536,0.008644444247086843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,256,0.020706666840447318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,65536,0.06711377700169881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,2560,0.014672888649834527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,8192,0.01094133324093289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,10240,0.012069332930776807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,128,0.00620977787507905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,2560,0.0151573336786694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,8192,0.015136000182893542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,768,0.014398222168286642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,768,0.018971555762820773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,768,0.009007111191749573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,3584,0.005771555420425203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,65536,0.65921598010593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,65536,0.09686755471759373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,1024,0.0035208890007601846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,256,0.002479111154874166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,128,0.0022195554855797025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,2048,0.011438222395049201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,5120,0.006639110959238476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,65536,0.06465599934260051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,4096,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,4096,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,16384,0.052388442887200244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,2560,0.009276444713274637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,6144,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,65536,0.12270932727389866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,3584,0.017236444685194228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,768,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,2048,0.020929776959949072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,7168,0.00998755544424057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,7168,0.04113066527578566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,1536,0.004234666625658671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,4096,0.007316444483068254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,65536,0.3243849012586805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,65536,0.08193777667151557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,16384,0.03933599922392104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,128,0.012925333446926542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,2560,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,10240,0.025620443953408137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,5120,0.009340444372759925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,16384,0.6032106611463759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,7168,0.09890133142471313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,12288,0.015788444214397006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,8192,0.010657777388890585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,1536,0.19354489114549425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,65536,0.3948800033993191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,16384,0.025283555189768474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,2560,0.01069777790043089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,2048,0.034472889370388456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,2048,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,6144,0.04587022132343716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,5120,0.11665244897206624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,3584,0.030018667380015056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,768,0.01513066722287072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,32,0.0017955555684036678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,16384,0.01648622254530589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,7168,0.07880889044867621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,7168,0.009668444593747457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,5120,0.004846222284767362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,64,0.011013333168294696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,5120,0.011001778145631155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,3584,0.004836444639497333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,512,0.014479999740918478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,7168,0.06422577963935004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,8192,0.030091553926467896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,4096,0.014764444695578681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,3584,0.015052444405025907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,65536,0.016069332758585613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,1024,0.009350222018029954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,16384,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,128,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,512,0.015455999308162265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,4096,0.018165333403481376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,65536,0.5122764375474718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,4096,0.014519999424616495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,32,0.0031662223239739737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,65536,0.07133955425686307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,64,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,1024,0.01737866633468204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,6144,0.013376889129479727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,16384,0.030159112479951646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,2048,0.010313777459992303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,1536,0.03889155719015334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,128,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,512,0.003428444473279847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,512,0.00833155545923445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,2048,0.01517511076397366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,1536,0.011900444825490316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,2048,0.021976000732845728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,64,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,2048,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,4096,0.02946044339074029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,1024,0.006433777925040986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,768,0.0069288888739215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,256,0.0028488888508743713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,128,0.010675555302037133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,512,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,128,0.005236444373925527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,4096,0.07374222411049737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,1536,0.006223111102978389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,7168,0.06856533553865221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,16384,0.02988622254795498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,256,0.014841778410805596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,12288,0.29895822207132977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,2048,0.004922666483455234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,8192,0.04308977723121643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,768,0.012979555461141797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,2048,0.01739377776781718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,7168,0.025083555115593806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,2560,0.04520622226927015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,8192,0.09665066666073269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,65536,2.4387440151638455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,12288,0.04184000028504265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,256,0.11001955800586277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,2560,0.02400355537732442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,6144,0.018547554810841877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,4096,0.013751111096805997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,4096,0.007679999702506595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,512,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,128,0.0022373333987262514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,2048,0.004217777815130022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,6144,0.0567128856976827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,16384,0.0838444431622823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,7168,0.02169422143035465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,1024,0.019914666811625164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,3072,0.030664887693193223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,12288,0.03885244329770406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,256,0.0027919999427265594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,4096,0.007961778177155389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,2048,0.08379289176728989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,3072,0.010650667051474253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,1536,0.008737777670224508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,6144,0.032207111517588295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,1536,0.025493333737055462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,128,0.0021679999513758556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,128,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,65536,0.5566435390048557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,16384,0.04433955417739021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,7168,0.019634667370054457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,2048,0.011034666664070554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,8192,0.03252799974547492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,7168,0.05221244361665514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,10240,0.012086222569147745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,10240,0.10649066501193577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,65536,0.12735022438897028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,512,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,768,0.0041742221348815495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,5120,0.12710844145880804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,3072,0.018571555614471436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,6144,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,10240,0.018208000395033095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,6144,0.11920621660020615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,12288,0.03931555483076308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,2048,0.006980444822046492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,768,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,16384,0.030238221089045208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,4096,0.027273777458402846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,6144,0.0038986665507157645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,3584,0.01055022246307797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,2048,0.1362231175104777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,1024,0.00590844452381134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,768,0.0032186667538351486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,12288,0.01683200067943997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,8192,0.014880888991885714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,256,0.001820444439848264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,1536,0.020453333854675293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,768,0.009400000174840292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,7168,0.015159999330838522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,10240,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,3584,0.024508444799317256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,16384,0.317068444357978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,128,0.016137777103318107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,10240,0.012440000143316058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,2048,0.017906667457686532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,3584,0.015325332681337992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,6144,0.01550044450494978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,3072,0.03603822324011061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,2560,0.008991999758614434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,7168,0.030864887767367895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,3072,0.007367110914654202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,8192,0.1416915522681342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,1024,0.044962667756610446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,128,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,1536,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,8192,0.011404444773991903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,7168,0.005372444374693765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,256,0.012646222280131446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,64,0.0017831110292010838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,1024,0.007602666815121968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,8192,0.019892444213231403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,256,0.017345777816242643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,3072,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,1536,0.014112888111008538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,4096,0.011039999624093374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,12288,0.02734222180313534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,10240,0.005579555614127054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,6144,0.024105777343114216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,8192,0.014659555421935188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,6144,0.01143555591503779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,64,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,3584,0.12293688456217448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,3072,0.027453333139419556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,2560,0.007270221908887227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,7168,0.016514667206340365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,10240,0.033193777004877724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,5120,0.017418666018380057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,2048,0.018544000056054857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,256,0.005752888818581899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,3072,0.011076444553004371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,65536,0.05736088752746582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,3584,0.007584888901975419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,10240,0.04227200150489807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,2048,0.006520000182920032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,5120,0.018187556001875136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,8192,0.031288888719346784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,6144,0.009385777844323052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,16384,0.033562666840023465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,12288,0.03212711215019226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,256,0.010022222167915767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,8192,0.06910844643910725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,512,0.013738666971524557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,768,0.006917333437336816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,5120,0.014498665928840637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,2560,0.013774221969975365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,12288,0.04239911172125074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,8192,0.011364444262451597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,1024,0.015848888291252982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,5120,0.07819555865393744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,768,0.019909333851602342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,65536,0.1261920001771715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,2560,0.28764179017808705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,5120,0.016878222425778706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,65536,0.06100177764892578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,4096,0.008270222279760573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,10240,0.015999999311235216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,2048,0.006873777757088344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,65536,0.1522995498445299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,10240,0.012015110916561551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,2048,0.014243554737832812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,10240,0.6072053379482693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,7168,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,4096,0.04244622257020739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,1024,0.021988444858127173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,768,0.01995111174053616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,7168,0.012068444656001197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,65536,0.0735351112153795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,16384,0.055389334758122764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,5120,0.01830844415558709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,768,0.014827556080288358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,3072,0.009391999906963771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,768,0.005987555616431766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,1024,0.0038702223036024305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,2048,0.003912888881233004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,128,0.004907555464241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,10240,0.031847112708621554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,256,0.011036444041464063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,7168,0.008947555389669206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,16384,0.14199111196729872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,4096,0.12050133281283909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,5120,0.01161688897344801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,1536,0.003536888708670934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,5120,0.010318222145239512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,4096,0.010994666980372535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,3584,0.028452446063359577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,2560,0.011342222491900126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,6144,0.020250666472646926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,10240,0.032271110349231295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,1024,0.0063271112740039825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,128,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,256,0.0034595556143257353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,5120,0.07787822352515326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,5120,0.015463999576038785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,6144,0.02195555633968777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,1536,0.012159110771285163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,7168,0.27717600928412545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,10240,0.03236888845761617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,64,0.0052782222628593445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,6144,0.02952444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,7168,0.010021333065297868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,2048,0.005979555762476391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,2560,0.03603733248180813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,1536,0.0062871111763848206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,16384,0.017183999220530193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,2048,0.01819022165404426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,2048,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,128,0.010816888676749336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,4096,0.012070222033394707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,8192,0.023028444912698533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,1536,0.015811555915408664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,128,0.02961155441072252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,12288,0.30580798784891766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,4096,0.04125866625044081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,5120,0.02313066687848833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,8192,0.04359644320276049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,768,0.018421333697107103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,64,0.001822222231162919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,8192,0.012159999873903064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,3072,0.008016000191370646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,1536,0.010334221853150262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,3584,0.022991110881169636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,8192,0.004486222234037188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,3584,0.010629333555698395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,128,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,6144,0.017855111095640395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,1536,0.06871377759509616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,65536,0.42136266496446395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,10240,0.012357333468066322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,16384,0.005958222266700532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,4096,0.01420177850458357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,2048,0.007327111230956183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,64,0.005606222069925732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,32,0.0039022221333450745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,2048,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,8192,0.01886666648917728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,1024,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,3584,0.016397333807415433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,8192,0.010718222293588849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,8192,0.01835555500454373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,4096,0.026239110363854304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,128,0.0019164443429973389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,10240,0.026168000366952684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,2560,0.00813422186507119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,65536,0.05263911022080315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,2048,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,7168,0.0103493332862854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,2560,0.021951110826598272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,256,0.0021555556191338436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,5120,0.009372444616423713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,3072,0.03995111253526475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,10240,0.016427555018001132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,3584,0.019949333535300363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,12288,0.018509333332379658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,12288,0.017813333206706576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,2048,0.006575111299753189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,7168,0.009975999593734741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,1024,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,16384,0.017314665847354464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,2560,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,4096,0.030624889665179785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,8192,0.022053332792388067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,6144,0.07997066444820829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,3072,0.021917333205540974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,7168,0.0226275556617313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,128,0.005771555420425203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,6144,0.018944889307022095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,6144,0.012348444097571902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,768,0.01797422269980113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,1536,0.006304889089531369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,10240,0.08080710967381795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,16384,0.02106399999724494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,1024,0.005879111174080107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,256,0.01514844430817498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,7168,0.010041777458455827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,12288,0.020021332634819877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,2560,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,5120,0.12957866986592612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,3072,0.006979555719428592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,3072,0.004906666775544484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,512,0.006682666639486949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,512,0.01237422227859497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,768,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,10240,0.01958044370015462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,3072,0.014107555150985718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,768,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,3072,0.008106666306654612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,5120,0.1799022224214342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,8192,0.009992889232105678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,3584,0.011692444483439127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,3584,0.017177777157889474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,64,0.0041786668201287585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,256,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,32,0.00554044461912579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,16384,0.10473155313067967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,65536,0.9456941816541883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,4096,0.01071466671095954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,2048,0.00424888895617591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,512,0.01461866663561927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,64,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,512,0.006904888898134232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,16384,0.01583377851380242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,7168,0.012889777620633444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,12288,0.014428445034556918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,12288,0.2853608926137288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,4096,0.008328888979223039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,4096,0.010079999764760336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,32,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,32,0.002491555487116178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,64,0.01683111157682207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,5120,0.008566222257084316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,1024,0.015434667468070984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,16384,0.01751377847459581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,6144,0.00923911150958803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,5120,0.008619555168681675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,16384,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,6144,0.009337777892748514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,65536,0.19287644492255318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,2560,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,2048,0.006554666492674086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,3072,0.01826400061448415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,256,0.002151111140847206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,10240,0.036687112516827054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,16384,0.49115557140774196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,1024,0.0034755555291970572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,3584,0.015114666687117683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,32,0.04962044292026096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,7168,0.04599110947714912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,16384,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,2560,0.0074346669846110875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,65536,0.1292897727754381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,3072,0.009704888694816166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,512,0.0232924438185162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,5120,0.01607377827167511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,12288,0.12735111183590359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,10240,0.017995556195576985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,768,0.0032791110376516977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,7168,0.004874666531880696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,32,0.0017928888814316855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,1536,0.021253334151373968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,10240,0.01718311177359687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,1024,0.0038337777886125776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,512,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,3072,0.018168000711335074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,16384,0.03807199994723002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,12288,0.018576888574494254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,7168,0.012723555167516073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,2560,0.02863022353914049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,2560,0.007236444287829929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,3072,0.029389331738154095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,128,0.002190222135848469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,3584,0.017166222135225933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,3072,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,2048,0.016862221890025668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,65536,0.3072897858089871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,768,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,12288,0.014088888963063559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,768,0.010990222295125326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,3072,0.020418667131000094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,2560,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,12288,0.02557777696185642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,10240,0.11267466015285915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,256,0.002509333400262727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,8192,0.14791022406684026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,10240,0.07282044490178426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,512,0.0033413333197434745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,16384,0.052107555998696216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,5120,0.009003555609120263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,2560,0.1554186609056261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,64,0.001828444500764211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,16384,0.013539555172125498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,3584,0.019564444820086162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,2048,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,2560,0.027441778116756018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,3584,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,7168,0.01328888866636488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,16384,0.022501332892311945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,2048,0.006638222270541721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,16384,0.10304622517691718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,2048,0.0042408886882993905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,8192,0.027890665663613215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,768,0.016840888394249808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,64,0.005229333208666907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,1024,0.0191928893327713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,65536,0.07149333424038358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,5120,0.09327466620339288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,1536,0.022555554906527202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,64,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,2048,0.011354666617181567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,65536,0.014905777242448596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,2048,0.00870488915178511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,2560,0.016007110476493835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,16384,0.019354666272799175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,64,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,2560,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,6144,0.017826666434605915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,64,0.016533333394262526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,4096,0.031529777579837374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,256,0.01648355523745219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,3072,0.15836000442504883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,8192,0.03058755397796631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,65536,0.07669688595665826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,3584,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,512,0.007114667031500075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,16384,0.029586666160159644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,3584,0.013170666992664337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,512,0.0022320000247822865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,2048,0.09723377890057033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,8192,0.032864888509114586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,128,0.005320000151793162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,4096,0.014316444595654806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,4096,0.04598844382497999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,128,0.002528888897763358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,12288,0.021232000655598108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,12288,0.00720355576939053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,10240,0.07875021961000231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,3072,0.0075937774446275495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,6144,0.01145333300034205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,12288,0.05374044511053297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,128,0.005221333354711533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,512,0.013425778183672162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,768,0.0034604445099830627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,65536,2.076491461859809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,1024,0.006529777828190062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,3072,0.010479110810491772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,1024,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,64,0.01646222174167633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,6144,0.017856000198258292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,12288,0.01756000022093455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,1024,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,12288,0.02365866634580824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,3072,0.004563555535342959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,1536,0.009308444129096137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,512,0.02086488902568817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,7168,0.04148799843258328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,10240,0.03990755478541056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,12288,0.014524444937705994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,2048,0.016919111212094624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,256,0.014832889040311178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,3072,0.007923555870850882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,128,0.02832177612516615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,512,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,4096,0.016192000773217943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,256,0.010999110837777456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,2560,0.010043555663691627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,128,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,1024,0.003169777699642711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,128,0.005950222412745158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,7168,0.015360888507631091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,8192,0.8677208688524035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,65536,0.07006488906012641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,12288,0.021312000022994146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,7168,0.027987556325064764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,2560,0.009843555589516958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,768,0.01774488886197408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,8192,0.014137778017255994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,2560,0.5559848679436578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,128,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,8192,0.01072088877360026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,3072,0.013346666263209449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,1024,0.005950222412745158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,1024,0.003495111233658261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,5120,0.01222222215599484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,16384,1.4615039825439453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,2048,0.007612444460391998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,1536,0.008110221889283922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,10240,0.02663822140958574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,6144,0.011950222154458364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,2048,0.023560888237423364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,5120,0.01418133411142561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,5120,0.021465778350830078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,10240,0.012686221963829465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,12288,0.015896888242827523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,256,0.0021253333737452826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,3584,0.15023199717203775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,6144,0.04569066564242045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,1536,0.00620888877246115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,8192,0.010084444450007545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,7168,0.3088497850630018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,3584,0.024093333217832778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,16384,0.046309332052866616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,8192,0.015431111057599386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,4096,0.08261866701973809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,3072,0.013809777796268463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,1024,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,256,0.09809599982367621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,16384,0.025786666406525507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,1024,0.014465777410401238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,2560,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,64,0.003924444317817688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,3584,0.00903911143541336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,128,0.012634666429625617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,65536,0.08759822448094685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,5120,0.01791555517249637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,7168,0.03048444456524319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,65536,0.21705867184533012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,7168,0.24302755461798775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,65536,0.03518666823705038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,3584,0.007751999629868402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,2048,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,12288,0.020610666937298246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,768,0.018745778335465323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,4096,0.00903555585278405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,8192,0.14247643947601318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,7168,0.0086666668454806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,768,0.005836444596449534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,16384,0.01145066652033064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,1024,0.06224711073769463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,16384,0.04319555560747782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,128,0.006217777729034424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,1536,0.009304888546466827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,64,0.006335999816656113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,3584,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,2560,0.004921777794758479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,3584,0.020558221472634208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,65536,0.5956577724880642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,16384,0.022543110781245764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,6144,0.03374400072627597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,1024,0.00328533330725299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,65536,0.08329421944088407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,128,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,8192,0.014351111319330005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,2560,0.023403555154800415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,4096,0.007718222008811102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,12288,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,3584,0.27079910702175564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,10240,0.022076444493399724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,5120,0.10627733336554633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,3072,0.01037688884470198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,768,0.0033315554675128725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,2048,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,10240,0.03218933277659946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,65536,0.06564444303512573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,10240,0.37794311841328937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,256,0.009396444592210982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,65536,0.14821599589453802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,2048,0.015649777319696214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,8192,0.02405244443151686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,512,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,12288,0.019839111301634047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,256,0.01800355480776893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,32,0.010461333725187512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,65536,0.0765297810236613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,8192,0.022907555103302002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,768,0.005666666560702854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,1536,0.003551111039188173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,2560,0.028074665202034846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,8192,0.023149333066410486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,8192,0.09171111053890652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,3072,0.08698488606346978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,7168,0.15583554903666177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,1024,0.024937777055634394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,5120,0.01953066719902886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,768,0.03528977764977349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,10240,0.015775110986497667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,3072,0.04518933428658379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,7168,0.07943644126256307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,3584,0.013119999733236102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,6144,0.017507554756270517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,16384,0.0600222216712104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,32,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,768,0.002842666581273079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,4096,0.020642666353119742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,2560,0.007104888558387756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,1536,0.029106666644414265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,2048,0.006912888752089606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,128,0.0018968888454967076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,3584,0.033077332708570696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,768,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,8192,0.020263110597928364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,768,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,3072,0.013467555244763693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,16384,0.7739689085218641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,1024,0.010444444086816577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,1024,0.015196444259749519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,64,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,12288,0.016506666938463848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,3584,0.08096355862087674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,5120,0.004912000149488449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,8192,0.013396444420019785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,4096,0.01484266585773892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,3072,0.026735999517970618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,7168,0.030118223693635728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,2560,0.06168711185455322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,10240,0.012416000167528788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,6144,0.009342222577995723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,65536,0.11738933457268609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,12288,0.013761777844693927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,64,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,5120,0.008654221892356873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,16384,0.016554666890038382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,768,0.0028595555987623003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,32,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,12288,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,3584,0.0612915555636088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,1536,0.028901331954532202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,64,0.0018168888572189545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,12288,0.018717333674430847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,3072,0.011071999867757162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,512,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,512,0.0036435553597079385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,1024,0.01585155559910668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,7168,0.02756266627046797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,768,0.009885333478450775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,16384,0.022252443763944838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,10240,0.13293866316477457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,2048,0.017161778277821012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,6144,0.034128000338872276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,3584,0.008292444050312042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,7168,0.022327999273935955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,65536,0.0878373318248325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,3584,0.1201288832558526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,768,0.0068195557428730865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,4096,0.013407111167907715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,3072,0.004595555365085602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,768,0.024076445235146418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,256,0.012998222476906247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,3584,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,3072,0.017992888887723286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,4096,0.011012444065676795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,64,0.20138577620188394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,256,0.0032906666811969546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,12288,0.01754044493039449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,2048,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,12288,0.11087199714448716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,8192,0.03389066788885329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,6144,0.00962577760219574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,10240,0.03813955518934462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,256,0.005396444350481033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,3584,0.007255111303594377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,1024,0.00600000015563435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,256,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,10240,0.015250666273964776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,2560,0.01479288935661316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,12288,0.0455057786570655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,10240,0.03288355469703674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,32,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,128,0.00184888889392217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,16384,0.053745776414871216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,16384,0.0052284445199701525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,5120,0.03051999873585171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,7168,0.035100443495644465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,8192,0.014800889624489678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,12288,0.020580444071027968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,3584,0.016338666280110676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,512,0.006055111272467508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,256,0.0028764444092909494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,1536,0.024272888898849487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,5120,0.01370666672786077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,10240,0.029550221231248643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,1024,0.003152888889114062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,3072,0.007654222349325816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,64,0.006411555740568373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,2560,0.018629334039158292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,128,0.013443555268976422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,256,0.018251554833518136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,3584,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,3072,0.024173332585228816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,3072,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,768,0.017335111896197002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,10240,0.027112000518374976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,3072,0.022960889670583937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,2048,0.004848000076082018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,10240,0.05292444427808126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,512,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,16384,0.028425776296191748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,128,0.0018791111393107309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,10240,0.01646933290693495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,2560,0.01481244464715322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,3584,0.022898667388492163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,768,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,8192,0.04942311180962456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,512,0.007264888948864407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,1024,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,768,0.013435555828942193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,768,0.005678222411208683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,2048,0.038450668255488075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,16384,0.0406951109568278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,64,0.0052595556610160405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,5120,0.017539555827776592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,12288,0.07633511225382487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,1024,0.006772444479995304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,128,0.010094222095277574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,7168,0.02015999952952067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,12288,0.02910311023394267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,512,0.0362684428691864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,4096,0.04885066548983256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,2048,0.005527110977305307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,65536,0.16639822059207493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,3584,0.6409813563028971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,3584,0.00796088907453749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,10240,0.01867555578549703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,10240,0.025007999605602686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,512,0.0028106667515304354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,3584,0.03137066629197862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,64,0.006051555689838197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,512,0.04862844281726413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,128,0.019902222686343722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,3584,0.007835555407736037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,12288,0.018767999278174508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,1536,0.009190222455395592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,2560,0.024859555893474158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,8192,0.012801777985360889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,6144,0.010061333576838175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,2048,0.0452497767077552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,7168,0.02350844442844391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,16384,0.023024888502226934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,16384,0.013733333183659447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,8192,0.38978666729397243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,10240,0.017887999614079792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,3072,0.00962222201956643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,4096,0.008454221818182204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,8192,0.010556444525718689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,7168,0.020207999481095206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,16384,0.030810667408837214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,768,0.017493334081437852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,7168,0.009642666412724389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,4096,0.01984177695380317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,128,0.006314666734801398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,7168,0.032122668292787336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,10240,0.008963555925422246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,512,0.0496346652507782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,32,0.0019742221468024785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,6144,0.006222222414281633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,128,0.010332444475756751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,1536,0.004893333133724001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,5120,0.012127999630239276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,5120,0.02572088936964671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,2560,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,64,0.005716444303592046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,128,0.013439999686347114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,64,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,65536,0.20095554987589517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,10240,0.012715555727481842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,3584,0.023352000448438857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,3072,0.010138666464222802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,256,0.009720889230569204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,12288,0.018904889623324077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,2048,0.0075937774446275495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,2048,0.015844444433848064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,1024,0.007644444704055786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,512,0.014490667316648694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,3584,0.016846223009957206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,4096,0.007965332931942409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,32,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,12288,0.06489777565002441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,32,0.01683200067943997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,2560,0.0070222218831380205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,3584,0.022026666336589392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,10240,0.01702133317788442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,12288,0.018612444400787354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,128,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,10240,0.044844445255067616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,16384,0.36517243915134007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,3584,0.02791022260983785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,6144,0.009569777382744683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,8192,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,1536,0.028796444336573284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,65536,4.68867450290256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,7168,0.022506665852334764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,2048,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,2560,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,128,0.008272888759771982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,128,0.015400000744395785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,32,0.0052168890833854675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,32,0.0069217777086628815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,10240,0.006474666711356904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,2560,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,3584,0.020671111014154222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,512,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,2560,0.007072889142566257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,512,0.0041706665522522396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,4096,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,3584,0.011576000187132092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,1024,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,16384,0.019679110911157396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,7168,0.022011554903454248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,3072,0.38899911774529355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,1536,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,1536,0.015142222245534262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,512,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,1536,0.020540444387329947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,768,0.009591110878520543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,512,0.0110871113008923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,128,0.03080444534619649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,5120,0.011704888608720569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,2560,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,5120,0.11910755104488796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,4096,0.012426666915416718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,7168,0.019316444794336956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,256,0.015141333142916361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,1024,0.006593777901596493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,64,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,5120,0.14492177963256836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,256,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,12288,0.036764442920684814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,65536,0.257518212000529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,8192,0.05886755386988322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,12288,0.018397332893477548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,2048,0.006545777950021956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,5120,0.10415377881791855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,16384,0.06241599718729655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,256,0.015488889482286243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,2560,0.017496888836224873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,1024,0.0028737777223189673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,1024,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,5120,0.02546311087078518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,3584,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,2560,0.035176889763938055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,3584,0.11481332778930664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,6144,0.06056888898213705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,10240,0.04490400022930569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,10240,0.07634311252170138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,10240,0.03372800019052293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,4096,0.025405334101782903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,1024,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,1536,0.004219555606444676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,2048,0.2413262261284722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,2560,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,65536,0.06609244479073419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,3584,0.019932443896929424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,2048,0.02232622272438473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,65536,0.023787554767396715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,3072,0.010943111446168689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,3072,0.042098667886522084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,1536,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,3072,0.017321777012613084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,256,0.006530666516886816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,256,0.0021555556191338436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,5120,0.008986666798591614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,12288,0.30838134553697377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,1024,0.007779555188284979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,128,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,2048,0.011377777490350934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,128,0.0021199999998013177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,8192,0.014922666880819531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,2560,0.006706666615274217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,256,0.0021377777059872947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,3584,0.006223111102978389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,10240,0.015264888604482016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,7168,0.04380177789264255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,65536,0.06864711311128405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,16384,0.04459199971622891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,3584,0.01368622233470281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,3584,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,3072,0.007248889240953658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,512,0.007288888924651676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,8192,0.07853955692715116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,2560,0.004888888862397936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,12288,0.03417333298259311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,65536,0.058548443847232394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,32,0.0038373333712418876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,2048,0.009946666657924652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,8192,0.014492443866199918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,512,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,5120,0.018920888503392536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,3584,0.007637333538797166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,4096,0.008983111215962304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,32,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,7168,0.011757333245542316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,10240,0.01564088960488637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,4096,0.02532888948917389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,4096,0.008817777865462834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,2048,0.06374400191836886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,3584,0.022571555442280237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,2048,0.019976889093716938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,3072,0.009121778110663096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,128,0.007277333074145847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,2560,0.10220888588163589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,256,0.011702222128709158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,16384,0.02534755567709605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,768,0.0028293333533737394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,768,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,8192,0.025303110480308533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,16384,0.04086933202213711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,65536,0.05103111267089844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,3072,0.007269333634111617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,7168,0.027514666318893433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,3072,0.005877333382765452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,7168,0.03769866625467936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,512,0.01313066648112403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,1024,0.014304000470373364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,16384,0.016525333126386006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,8192,0.03770577907562256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,6144,0.005223111146026188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,4096,0.013399110900031196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,768,0.0028444443725877335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,8192,0.0107004443804423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,10240,0.03668977816899618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,16384,0.045943111181259155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,3072,0.008338666624493068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,8192,0.04588355620702108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,5120,0.03666399916013082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,7168,0.02702666653527154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,3072,0.009024000002278222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,256,0.015114666687117683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,64,0.0028124445428450904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,12288,0.014383110735151501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,768,0.008236444658703273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,128,0.014499555031458536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,128,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,16384,0.623944017622206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,256,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,10240,0.023660444551044043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,1536,0.021631111701329548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,1024,0.016878222425778706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,512,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,2560,0.010068444742096795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,4096,0.026769777139027912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,256,0.002520888836847411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,32,0.00217333332531982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,2048,0.16639555825127497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,32,0.003157333367400699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,10240,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,512,0.019941333267423842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,16384,0.07258577479256524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,8192,0.013108444710572561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,65536,0.08715111017227173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,12288,0.012718222207493253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,256,0.0090515555606948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,12288,0.05890666776233249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,128,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,10240,0.026699556244744196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,768,0.0025271111064487034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,128,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,3584,0.011141333315107556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,2560,0.009346666435400644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,7168,0.010305778019958073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,1024,0.017575111654069688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,3584,0.007589333587222629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,512,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,32,0.01146488885084788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,3584,0.013763555222087435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,12288,0.01924977699915568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,65536,0.2671288914150662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,64,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,6144,0.01961688862906562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,3072,0.004600000050332812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,768,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,10240,0.029751112063725788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,16384,0.14077688588036433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,10240,0.012827555338541666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,2560,0.007242666350470648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,512,0.009521777431170145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,3072,0.009941333697901832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,768,0.014140443669425117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,256,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,16384,0.16826311747233072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,768,0.002805333377586471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,16384,0.09699022107654148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,16384,0.015829333000712924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,128,0.01421511173248291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,768,0.007320000065697565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,1024,0.0032088889016045462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,128,0.01096088853147295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,8192,0.022330666581789654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,2560,0.016511999898486666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,2048,0.007083555890454187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,1536,0.00423911131090588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,2048,0.025632888078689575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,3072,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,10240,0.013443555268976422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,12288,0.010196444061067369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,65536,0.4262293444739447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,12288,0.037323554356892906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,16384,0.14673066139221191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,5120,0.004952888935804367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,128,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,1536,0.021839999490314062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,5120,0.010381333529949188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,12288,0.25520088937547475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,7168,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,1024,0.0029120000286234748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,10240,0.04333244429694282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,10240,0.0503768887784746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,256,0.006226666685607698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,512,0.0024604443460702896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,7168,0.1907120015886095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,1536,0.008055110772450766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,4096,0.011349333657158745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,10240,0.03587022092607286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,3072,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,3584,0.04787733488612705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,3584,0.007958222594526079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,16384,0.025573333104451496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,32,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,6144,0.009289777941173976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,64,0.00184888889392217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,256,0.015491555134455362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,7168,0.01219555570019616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,5120,0.02121066715982225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,16384,0.00922400007645289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,10240,0.020590222544140287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,6144,0.018211555149820115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,1536,0.030807998445298936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,3072,0.011666666302416058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,3584,0.008014221986134848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,12288,0.055418666866090566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,256,0.012072000238630505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,2560,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,768,0.019313777486483257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,2560,0.004873777843183941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,256,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,256,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,32,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,10240,0.061520000298817955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,512,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,12288,0.10739288727442424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,1536,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,5120,0.009008888569143083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,1536,0.021351999706692163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,3584,0.010028444230556488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,128,0.020119110743204754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,16384,0.07843733496136136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,256,0.0026026666164398193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,4096,0.5108702447679307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,64,0.030447999636332195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,1024,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,2048,0.02385155525472429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,2560,0.029104888439178467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,3072,0.015478221906556023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,1536,0.008625778059164682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,512,0.010679110884666443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,2560,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,10240,0.047431111335754395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,10240,0.14198488659328884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,65536,0.1502293348312378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,12288,0.0317884451813168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,1024,0.021331555313534204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,65536,0.02756622268093957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,512,0.0028000000036425064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,3072,0.025666667355431452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,3584,0.015839111473825242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,512,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,8192,0.014484445254007975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,8192,0.020431111256281536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,512,0.011730666789743634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,512,0.04660177893108792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,2560,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,512,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,3584,0.017665778597195942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,16384,0.01719911065366533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,768,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,768,0.0069884442620807225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,1024,0.005953777581453323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,2560,0.022904000348514978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,10240,0.12692799833085802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,10240,0.039800001515282526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,1024,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,16384,0.4044817818535699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,3072,0.17309510707855225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,32,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,128,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,2048,0.007986666427718269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,512,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,256,0.0032142222755485107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,128,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,4096,0.008631111019187504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,3584,0.04211733407444424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,3584,0.012016889121797351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,768,0.0031893334041039148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,1536,0.008633777499198914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,768,0.006416888700591193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,32,0.005375111268626318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,3072,0.004591111093759537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,128,0.003567999849716822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,65536,0.09161066346698338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,16384,0.04461510976155599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,6144,0.05614488654666477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,512,0.005909333212508096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,3584,0.021956443786621094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,65536,0.10305155648125543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,64,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,4096,0.015824000040690105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,16384,0.01791111131509145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,2048,0.08640799919764201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,1024,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,12288,0.038387556870778404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,12288,0.015812445018026564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,6144,0.009498666558000777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,256,0.014827556080288358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,2048,0.015792000624868605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,4096,0.010640888578361936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,8192,0.02258488867017958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,128,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,5120,0.0107022225856781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,5120,0.016840888394249808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,10240,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,3072,0.03142044279310439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,128,0.0035128887328836652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,2048,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,256,0.015802666544914246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,65536,0.1548480060365465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,3584,0.037003556887308754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,256,0.0021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,768,0.0031511110977994073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,1536,0.003875555677546395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,1536,0.07015111049016316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,64,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,4096,3.032350116305881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,12288,0.017264000243610807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,6144,0.019166222876972623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,512,0.012071111136012606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,768,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,7168,0.9175937440660266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,65536,0.08751999669604832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,1536,0.016383111476898193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,2560,0.01625688870747884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,65536,0.029374221960703533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,2048,0.006144888699054718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,5120,0.058396445380316846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,4096,0.01086400035354826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,10240,0.023670223024156358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,2048,0.007994666695594788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,6144,0.012464000119103326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,3072,0.04557333389918009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,128,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,32,0.001822222231162919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,4096,0.00943999985853831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,8192,0.048672888014051646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,2048,0.010428444378905825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,256,0.012296888563368054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,2560,0.019671110643280875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,2560,0.006769777586062749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,4096,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,128,0.01579466627703773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,7168,0.06650933292176989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,512,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,6144,0.009391110804345872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,65536,1.174001799689399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,4096,0.013047999805874295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,3072,0.011031111081441244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,65536,0.12783111466301814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,16384,0.031762666172451444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,16384,0.02203911046187083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,32,0.002159111201763153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,3072,0.010753778119881948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,768,0.021920000513394673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,16384,0.006952889263629913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,1024,0.021714667479197185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,6144,0.00941422250535753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,256,0.008599999878141616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,8192,0.07911022504170735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,2048,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,8192,0.0469084448284573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,2048,0.028965334097544353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,3584,0.008973333570692275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,32,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,16384,0.022042666872342426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,8192,0.047242667939927846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,5120,0.012326222327020435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,8192,0.02320266597800785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,768,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,3584,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,768,0.005620444400442972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,2048,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,7168,0.04571200079388089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,1024,0.009312888814343346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,7168,0.01850400037235684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,2560,0.01921333372592926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,256,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,10240,0.018799111247062683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,128,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,512,0.004190222256713443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,6144,0.1339555581410726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,65536,0.020029332902696397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,3072,1.3999884923299153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,768,0.007600888609886169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,5120,0.03867111272282071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,2048,0.01552177800072564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,512,0.008488000267081791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,512,0.018911111685964797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,8192,0.02641600039270189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,8192,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,128,0.06972889105478923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,32,0.0028888889484935333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,7168,0.01033333357837465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,10240,0.0210515558719635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,12288,0.030528889762030706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,768,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,1024,0.006598222172922558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,3584,0.007785778078767989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,3072,0.007632888853549957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,3072,0.016859556237856548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,6144,0.019928000039524503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,768,0.01755466726091173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,128,0.010001777774757808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,32,0.005414222263627582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,2560,0.08136177725262113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,6144,0.026553778184784785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,16384,0.016169778174824186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,8192,0.011576889289749993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,8192,0.01292799992693795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,6144,0.026727111803160772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,2048,0.006647999915811751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,2048,0.006591111007663939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,8192,0.015287111202875773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,2048,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,7168,0.04690488841798571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,16384,0.0407626661989424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,12288,0.02597066594494714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,7168,0.01496266656451755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,3584,0.007702222300900354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,10240,0.17896000544230142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,2560,0.007275555696752336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,3072,0.013787555197874704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,128,0.006672888994216919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,768,0.007649777664078607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,65536,0.18525066640641955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,32,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,2560,0.03561066587766012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,3072,0.008295111358165741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,5120,0.0192266669538286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,128,0.005220444252093633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,5120,0.011995555626021491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,8192,0.35431377092997235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,256,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,3584,0.014934221903483072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,128,0.0053671110007498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,512,0.007794666621420119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,7168,0.019911999503771465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,16384,0.05603733327653673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,768,0.005961777849329843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,4096,0.016565332810084026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,10240,0.049060444037119545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,64,0.0018622221218215094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,12288,0.0736231141620212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,12288,0.01034400032626258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,10240,0.03704977697796292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,8192,0.020591111646758187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,512,0.0042160000238153665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,5120,0.060476442178090416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,3072,0.06810044580035739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,256,0.002540444541308615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,4096,0.03612355391184489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,65536,0.0868897769186232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,6144,0.010991999672518836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,4096,0.016082665986484952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,2048,0.004513777792453766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,12288,0.04901688959863451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,10240,0.02872177627351549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,10240,0.3830897808074951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,4096,0.016525333126386006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,6144,0.012423111332787408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,64,0.0018053334206342697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,1536,0.012877333495352002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,5120,0.02604533400800493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,12288,0.011375111010339526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,16384,0.016179554992251925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,2560,0.007259555160999298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,65536,0.02200977835390303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,7168,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,1536,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,10240,0.010404444403118556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,5120,0.06875733534495036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,7168,0.17627288235558403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,2048,0.01688799924320645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,2560,0.004550222307443619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,32,0.006703111032644908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,12288,0.04623466730117798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,2048,0.010524444282054901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,3072,0.03589955634540982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,16384,0.015061333775520325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,7168,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,8192,0.012286221815480126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,7168,0.0414684447977278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,2560,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,32,0.002493333278430833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,2048,0.00869688888390859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,32,0.0019199999256266488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,8192,0.09860089090135361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,7168,0.044759111271964185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,768,0.009306666751702627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,256,0.0052968888647026485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,1024,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,2560,0.01737866633468204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,2560,0.011386666860845355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,10240,0.016927111479971144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,64,0.0028497777465316984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,12288,0.01568711135122511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,7168,0.051631109582053296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,512,0.002493333278430833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,8192,0.020571556356218126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,8192,0.01370488852262497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,1024,0.0034666665726237824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,768,0.005623111294375525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,3072,0.008995555341243744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,32,0.00797155582242542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,12288,2.320503022935655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,65536,0.24070845709906685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,32,0.0028444443725877335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,3584,0.00793866647614373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,16384,0.5174542003207737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,12288,0.02495999965402815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,8192,0.012105777859687805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,256,0.0023315555105606713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,6144,0.00488800017370118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,65536,0.06613955232832167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,7168,1.2594480514526367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,128,0.0027946666296985415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,32,0.0074826669361856245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,7168,0.01812266641192966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,512,0.015841777125994366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,1536,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,128,0.002830222249031067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,10240,0.018373333745532565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,128,0.005468444277842839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,1536,0.006559111177921295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,64,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,1536,0.07587999767727323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,256,0.008309333688682979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,256,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,4096,0.00866844422287411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,65536,0.08141066630681355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,128,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,1024,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,3584,0.015817777978049386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,1536,0.003936000168323517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,64,0.013032889200581444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,3584,0.04492444462246365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,8192,0.018583110637134977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,16384,0.016152888536453247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,128,0.006292444550328785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,128,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,768,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,2048,0.02239199976126353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,10240,0.02520977788501316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,512,0.003992889076471329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,4096,0.01718488832314809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,8192,0.011457777685589261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,5120,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,256,0.005201777650250329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,6144,0.013930666777822705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,1536,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,2048,0.0063786668082078295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,64,0.0018417777286635505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,32,0.001842666624320878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,1024,0.0034924443397257063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,3584,0.016488000750541687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,1536,0.00665511108107037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,6144,0.13043377134535047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,7168,0.03664533297220866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,512,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,3072,0.018949333164427016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,5120,0.10746666457917954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,6144,0.01144088887506061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,12288,0.5668560134039985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,65536,0.15630933973524305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,5120,0.015499555402331881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,128,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,256,0.019458666443824768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,4096,0.01641244524055057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,768,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,65536,0.22635555267333984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,8192,0.01092266705301073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,3584,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,128,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,8192,0.021985777550273474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,1536,0.014925332532988654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,6144,0.012190221912331052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,5120,0.011052444577217102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,5120,0.012062222593360476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,8192,0.01832266648610433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,12288,1.0905360115898979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,256,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,1024,0.02696977721320258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,10240,0.29031289948357475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,768,0.03973066806793213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,6144,0.013668444421556262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,8192,0.012063110868136087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,768,0.02402666707833608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,2048,0.10047733121448094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,256,0.005314666777849197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,1024,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,8192,0.014105778601434497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,128,0.019326221611764696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,65536,0.30613689952426487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,3072,0.034903112385008074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,256,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,512,0.0024826667375034755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,16384,0.03523555397987366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,3584,0.007727999654081132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,3072,0.007271111011505127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,256,0.006657777975002925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,2560,0.01408622165520986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,4096,0.06937511099709405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,12288,0.08453244633144802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,4096,0.047239998976389565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,512,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,4096,0.028990222348107234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,16384,0.06875822279188368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,32,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,5120,0.012848888834317526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,3072,0.004314666820896996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,16384,0.021570665968788996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,10240,0.04073866539531284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,512,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,2048,0.009155555731720394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,5120,0.0314631097846561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,3072,0.01032266683048672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,4096,0.08905777666303848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,12288,0.05489244394832187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,1024,0.10130755768881904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,32,0.001791999985774358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,1024,0.002926222152180142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,128,0.014750222365061441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,768,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,256,0.005712000032265981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,3072,0.008661333057615492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,3072,0.021976000732845728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,256,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,1024,0.006250666661394968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,5120,0.025796444879637823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,1024,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,768,0.004858666823969947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,4096,0.10316444767846002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,10240,0.011395555403497485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,5120,0.00961511085430781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,512,0.018553778529167175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,16384,0.03474222289191352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,65536,0.4596204227871365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,128,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,64,0.0017928888814316855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,1536,0.006950221955776215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,64,0.003919110943873723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,512,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,128,0.007637333538797166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,1536,0.006267555471923616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,1024,0.009388444324334463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,16384,0.022527111901177302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,512,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,8192,0.039545777771208025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,256,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,768,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,12288,0.020113777783181932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,3584,0.03180977702140808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,32,0.009033777647548253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,768,0.016979555288950603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,512,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,1536,0.006258666515350342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,512,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,65536,0.0691502226723565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,4096,0.01683999929163191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,3072,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,1024,0.006223111102978389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,1024,0.003539555602603488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,16384,0.07606133487489489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,768,0.024971556332376268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,7168,0.010002666877375709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,7168,0.0177502218219969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,1536,0.032413333654403687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,3072,0.004569777597983678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,4096,0.026133333643277485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,5120,0.036679112248950534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,3584,0.008593777815500895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,7168,0.016894222961531747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,4096,0.0035964445107513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,4096,0.014169777433077494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,1024,0.006304889089531369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,256,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,512,0.07189866569307116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,1536,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,3072,0.0074480002125104266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,12288,0.013134222063753339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,128,0.01479111115137736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,128,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,1024,0.005631999837027655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,768,0.006243555496136348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,2560,0.0795653329955207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,128,0.013764444324705334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,7168,0.004957333207130432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,4096,0.014676445060306124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,2048,0.012293332980738746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,7168,0.06647021902932061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,128,0.0031786666562159858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,32,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,64,0.0029306666304667792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,16384,0.024108444650967915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,1024,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,768,0.003519999898142285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,768,0.10219289196862115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,65536,0.4929528766208225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,256,0.004855999930037392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,3584,0.017280888226297166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,64,0.005537777725193236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,1536,0.017322666115230985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,1536,0.00629066675901413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,8192,0.014191110928853353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,768,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,3584,0.01906222270594703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,8192,0.17140532864464653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,2560,0.006976889239417181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,4096,0.08230400085449219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,16384,0.01762488815519545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,8192,0.01070577816830741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,8192,0.043284446001052856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,12288,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,3584,0.008549333446555668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,65536,0.22208533022138807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,3072,0.02367288867632548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,7168,0.020614221692085266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,65536,0.06836177905400594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,7168,0.010690666735172272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,3584,0.022282666630215112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,128,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,128,0.005347555710209741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,256,0.019288000133302476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,1536,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,3584,0.016414221790101793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,1536,0.0038808890514903595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,1024,0.012479999827014076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,6144,0.009107555780145857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,256,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,7168,0.017628444565667045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,5120,0.025992888543340895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,768,0.01888444523016612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,2048,0.017789334058761597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,768,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,64,0.005578666511509154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,1536,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,1024,0.041060444381501936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,6144,0.010390222072601318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,1536,0.003863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,256,0.0024746666765875286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,128,0.017162667380438913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,3072,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,16384,0.03445955448680454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,7168,0.01997866729895274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,1536,0.01633155511485206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,2048,0.01886399918132358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,2560,0.013362666798962487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,10240,0.04730222291416592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,65536,1.5985235638088648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,1024,0.006659555352396435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,128,0.016528889536857605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,512,0.008631111019187504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,256,0.005320000151793162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,4096,0.015598222613334656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,1536,0.00628977765639623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,768,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,768,0.0028435554769304064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,128,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,1536,0.006625777731339137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,65536,10.169160630967882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,32,0.009480000370078618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,16384,0.324452453189426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,65536,0.3856817881266276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,32,0.0018035556293196147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,1536,0.008344000412358178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,256,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,65536,0.3344568941328261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,5120,0.03477777706252204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,1536,0.021542222963439092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,16384,0.01716800034046173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,512,0.012717333104875354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,10240,0.09118400017420451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,16384,0.05506933397716946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,2560,0.01720711092154185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,6144,0.032953778902689614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,16384,0.1301893393198649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,5120,0.018459555175569322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,1024,0.03132355544302199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,512,0.0646853314505683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,12288,0.05639733208550347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,8192,0.015319999721315173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,4096,0.008271999657154083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,32,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,128,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,10240,0.01276533305644989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,10240,0.049703111251195274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,64,0.007784888976150089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,128,0.015118221441904703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,5120,0.008939555949634975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,1536,0.006310222049554189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,16384,0.020081778367360432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,16384,0.08300444152620104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,16384,0.07170755333370633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,3072,0.06911555263731214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,7168,0.004919111314747068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,5120,0.01611199975013733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,10240,0.015736889508035447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,32,0.0024666666156715816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,256,0.002191999927163124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,2048,0.017497777938842773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,3072,0.05503999855783251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,768,0.005535999933878581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,4096,0.024634665913052026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,256,0.0028364445186323593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,5120,0.02146044373512268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,10240,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,1024,0.02786933382352193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,5120,0.02365600069363912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,3584,0.008647999829716152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,8192,0.01074488874938753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,256,0.0025928889711697898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,2048,0.01293511109219657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,256,0.0021280000607172647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,64,0.005535999933878581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,4096,0.010668444136778513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,12288,0.01685244507259793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,4096,0.09769155581792195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,5120,0.015426667200194465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,16384,0.06449955701828003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,3584,0.006682666639486949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,128,0.005199111170238919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,8192,0.029049777322345312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,2048,0.011375111010339526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,8192,0.02574933403068119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,5120,0.1333671145968967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,128,0.014087110757827759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,4096,0.004513777792453766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,1536,0.08761333094702826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,1024,0.005224888937340842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,7168,0.010394666757848527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,768,0.005616888817813661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,512,0.04731200138727824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,10240,0.629827552371555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,2560,0.01372977760103014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,1024,0.008621333373917473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,7168,0.01238844460911221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,768,0.007191999918884701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,65536,0.12837866942087808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,5120,0.06484622425503202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,3584,0.015170666906568738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,512,0.017491555876202054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,3584,0.024439111351966858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,128,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,5120,0.012096000214417776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,1024,0.0028657778683635923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,1536,0.015551111764378019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,2560,0.01443733274936676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,2048,0.019828443725903828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,2048,0.015564444992277356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,16384,0.020672000116772123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,16384,0.005201777650250329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,2048,0.007115555306275685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,16384,0.01754044493039449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,256,0.015167999598715039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,7168,0.016571554872724745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,3584,0.009143999881214565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,65536,0.14173155360751682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,7168,0.009728888670603434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,16384,0.017419555120997958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,256,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,2560,0.034641779131359525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,10240,0.012048000262843238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,3584,0.02695200012789832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,768,0.009007111191749573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,16384,0.05029066734843784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,3072,0.010536889235178629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,2560,0.011755555868148804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,3584,0.04824799961513943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,256,0.01484444406297472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,768,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,128,0.009014222357008193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,4096,0.04737777842415703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,3072,0.023345778385798138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,10240,0.012229333321253458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,3584,0.03510133425394694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,3072,0.040748443868425153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,32,0.002495999965402815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,3584,0.027809778849283855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,6144,0.07744444741143121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,4096,0.024550222688251074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,1536,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,256,0.002537777854336633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,2560,0.007291555404663086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,2560,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,12288,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,8192,0.01147555559873581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,7168,0.025341333614455328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,7168,0.013391999734772576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,2560,0.01578044394652049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,2048,0.006924444602595435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,3072,0.01847111185391744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,2048,0.007483555210961237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,3584,0.05417244301901924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,768,0.02347733411524031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,256,0.008074666890833113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,3584,0.00757155567407608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,4096,0.13024622864193386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,3584,0.03639288743336996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,2560,0.00866844422287411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,2560,0.007242666350470648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,16384,0.012687111066447364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,6144,0.04138933287726508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,10240,0.01202311118443807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,7168,0.014296889305114746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,1024,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,65536,0.07818666431638929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,256,0.01737422247727712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,65536,0.06662488645977445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,1536,0.006963555183675554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,10240,0.014377777775128683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,3584,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,128,0.006507555643717448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,2560,0.01074222226937612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,65536,0.0723315543598599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,256,0.005221333354711533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,1024,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,12288,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,32,0.0018355554590622585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,5120,0.013735999663670858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,3072,0.07134755452473958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,10240,0.016570667425791424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,16384,0.05991911225848728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,5120,0.012748444245921241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,2560,0.007263110743628607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,16384,0.03393955694304572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,4096,0.008288000192907121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,5120,0.016711999972661335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,3584,0.02438755498992072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,6144,0.020563556088341605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,65536,0.37761155764261883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,6144,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,16384,0.0612693362765842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,512,0.01531644496652815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,2560,0.00903555585278405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,128,0.0024497778051429326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,1024,0.024010666542583044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,1024,0.014115555418862237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,2560,0.01479644411140018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,768,0.005983110931184556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,2560,0.017077333397335477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,64,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,256,0.005757333503829108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,5120,0.01889511115021176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,1024,0.03573244478967454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,3072,0.007676444119877285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,10240,0.08016800218158297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,768,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,32,0.0017964444640609955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,1536,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,64,0.0018008889423476325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,3072,0.009386666946940953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,10240,0.03708088729116652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,6144,0.012551999754375882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,4096,0.02502577834659153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,128,0.009672000176376766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,1536,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,256,0.006626666833957036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,2048,0.01943911115328471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,512,0.01994311147265964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,16384,0.019370666808552213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,256,0.016864000095261466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,3072,0.009329777624871995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,2048,0.01700711084736718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,8192,0.015604444675975375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,3584,0.20591999424828422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,10240,0.02883466747072008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,4096,0.008310221963458592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,32,0.006302222195598815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,65536,0.1055484414100647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,16384,0.17629422081841362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,1024,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,12288,0.017851556340853374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,65536,2.713808059692383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,2560,0.017518222332000732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,6144,0.028883556524912517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,2560,0.013797333670987023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,2048,0.019944889677895438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,3584,0.052200890249676175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,1536,0.02494488822089301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,3072,0.017602667212486267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,32,0.00685066646999783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,2560,0.010655110908879174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,128,0.006309333360857434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,8192,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,256,0.012094222009181976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,512,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,512,0.016141333513789706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,2048,0.018230222993426852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,256,0.005195555587609609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,6144,0.030976888206270006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,16384,0.04307377669546339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,32,0.00553688903649648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,5120,0.028999110062917072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,64,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,1536,0.02018044392267863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,7168,0.009343110852771336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,32,0.006586666736337874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,3072,0.023061333431137934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,3072,0.011057777537239922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,8192,0.058851553334130176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,5120,0.03784711162249247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,2560,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,3584,0.010386666489972008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,768,0.007609777980380588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,4096,0.01865600049495697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,7168,0.06005244122611152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,1536,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,16384,0.024280889166726008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,1024,0.0031511110977994073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,6144,0.02218311031659444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,3584,0.008344000412358178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,65536,0.016202666693263583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,4096,0.008311111066076491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,4096,0.05631822347640991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,256,0.005606222069925732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,256,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,12288,0.08502311176723903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,128,0.005611555443869696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,12288,0.1675635576248169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,8192,0.015160888433456421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,6144,0.009719111025333405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,8192,0.021917333205540974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,6144,0.019868445065286424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,6144,0.09928800000084771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,10240,0.012730666332774691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,64,0.006613333192136552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,4096,0.007982222570313348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,768,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,16384,0.035438223017586604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,4096,0.029450668228997126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,4096,0.011056000159846412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,1536,0.006552000012662675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,128,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,512,0.005616888817813661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,64,0.006591111007663939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,1024,0.006575111299753189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,8192,0.05214222272237142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,6144,0.044975999328825206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,7168,0.00998933364947637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,12288,0.03286222285694546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,7168,0.010043555663691627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,12288,0.013049778011110095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,8192,0.01590488851070404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,128,0.015440889530711703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,10240,0.02573066618707445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,3584,0.07199377483791776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,768,0.015017777681350708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,7168,0.019762666689025033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,3072,0.17751288414001465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,2560,0.014807111687130399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,128,0.01407911048995124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,128,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,2048,0.004910222358173794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,1024,0.0032133333798911837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,2560,0.0203840004073249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,4096,0.008646222452322641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,32,0.002569777684079276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,3072,0.00923733330435223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,64,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,8192,0.047822223769293896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,32,0.0021911110315057966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,128,0.007610666255156199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,7168,0.6269529130723741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,2048,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,3072,0.006834666762087081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,6144,0.11564800474378799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,1536,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,65536,0.04939555459552341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,1024,0.02363288899262746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,10240,0.02230666743384467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,64,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,2048,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,1536,0.003875555677546395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,10240,0.032107555203967623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,4096,0.01308444473478529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,16384,0.02511288887924618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,128,0.06252800093756782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,65536,0.28287111388312447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,8192,0.22488800684611002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,256,0.002528888897763358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,3584,0.01074044406414032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,256,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,7168,0.013549333645237817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,3584,0.009368889033794403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,3072,0.009970666633711921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,6144,0.045851555135515004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,256,0.07062311304940118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,256,0.014948444234000312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,512,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,1536,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,3072,0.003911111089918348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,16384,0.03375822305679321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,12288,0.015325332681337992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,6144,0.011379555695586734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,2560,0.007975999679830339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,5120,0.01036622209681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,128,0.001785777716173066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,2048,0.008208889100286696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,128,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,32,0.015816888875431485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,1536,0.015503111812803479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,2048,0.004519999855094486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,1536,0.01719911065366533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,10240,0.10772889190249973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,2048,0.03683111071586609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,4096,0.0081386665503184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,512,0.005531555662552516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,2560,0.02749600013097127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,16384,0.01181688904762268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,65536,0.4381857713063558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,8192,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,256,0.0031884445084465873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,2560,0.03913244605064392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,3072,0.02070488863521152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,12288,0.01814844376511044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,16384,0.15956444210476345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,12288,0.019631110959582858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,256,0.07168000274234347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,512,0.00564533347884814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,3072,0.017723555366198223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,8192,0.010683555569913654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,65536,0.08003911044862536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,32,0.00908622228437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,1024,0.007306666837798224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,3584,0.011066666907734342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,6144,0.009367999931176504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,5120,0.0724053316646152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,8192,0.023333332604832117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,2560,0.07554222477806939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,65536,0.06737244129180908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,1536,0.012773333324326409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,10240,0.13019555144839817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,3072,0.01479466590616438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,256,0.007194666398896112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,1536,0.019292443990707397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,7168,0.02290488945113288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,512,0.0025271111064487034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,128,0.01721244388156467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,512,0.011357333097192975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,7168,0.012930666406949362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,6144,0.009025777379671732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,128,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,7168,0.042641778786977134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,128,0.0024968888610601425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,65536,0.17678044901953804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,2048,0.006506666541099548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,768,0.00249777775671747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,7168,0.022290666898091633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,16384,0.04860088891453213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,8192,0.0052737775776121355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,7168,0.013429332938459186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,2560,0.015115555789735583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,12288,0.43779910935295957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,768,0.016530666086408827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,4096,0.007205333146784041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,12288,0.0414720012081994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,3072,0.07719644572999743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,16384,0.04499377806981405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,16384,0.04522222280502319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,7168,0.018805333309703402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,2560,0.015299555328157214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,6144,0.00867555538813273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,1536,0.00925244473748737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,32,0.0014808888857563336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,4096,0.015956444872750174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,8192,0.013813333378897773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,1024,0.012313777373896705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,65536,0.10198399755689834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,7168,0.12245599428812663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,16384,0.018584888842370775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,1024,0.062111112806532115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,5120,0.009338666995366415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,5120,0.024805333879258897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,8192,0.01071466671095954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,10240,0.05375555488798353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,32,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,7168,0.011808888779746162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,4096,0.020936000678274367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,65536,0.8339760038587781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,1536,0.006227555374304454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,2048,0.008775999976529015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,256,0.005614222337802251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,16384,0.05077244506941902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,7168,0.010341333018408881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,4096,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,768,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,65536,0.8832560115390353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,2048,0.004597333156400257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,256,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,1024,0.025594666600227356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,10240,0.16724888483683267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,4096,0.0240346673462126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,5120,0.011482666763994427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,1024,0.026591110560629103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,512,0.013120888835854001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,3584,0.014497778481907315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,10240,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,128,0.011753777662913004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,128,0.02885955572128296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,7168,0.01254488858911726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,6144,0.017530666457282174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,256,0.0024097777075237697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,8192,0.09969155655966865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,3584,0.007681777907742395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,16384,0.030615998639000788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,1024,0.017086222767829895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,4096,0.03987110985649957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,7168,0.5186737908257378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,32,0.006331555545330048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,512,0.056193775600857206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,1536,0.024103111690945093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,1024,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,6144,0.005218666460778978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,32,0.010335110955768162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,4096,0.011704888608720569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,512,0.006566222343179915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,3584,0.008784000244405534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,6144,0.010950222611427307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,256,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,32,0.006613333192136552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,3584,0.007432888779375289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,8192,0.016004444824324716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,10240,0.010334221853150262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,4096,0.008328888979223039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,1536,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,768,0.005434666656785541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,1536,0.007253333098358578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,768,0.006622222148709827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,2048,0.04594044552909004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,16384,0.048799110783470996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,768,0.0038533334930737815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,12288,0.014677332507239448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,6144,0.01846311158604092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,32,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,3584,0.014134221606784396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,64,0.011092444260915121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,4096,0.010704889065689512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,12288,0.06257777743869357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,768,0.0411119990878635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,3584,0.01461155547036065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,2560,0.019498666127522785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,16384,0.05322755707634819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,12288,0.011238222320874533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,3072,0.008283555507659912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,10240,0.02232711182700263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,3584,0.01921866668595208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,16384,0.025804445147514343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,128,0.012070222033394707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,1024,0.014130666851997375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,16384,0.12808356020185682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,128,0.00655644428398874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,2560,0.00999644481473499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,2560,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,5120,0.02026844521363576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,32,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,6144,0.01144977741771274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,6144,0.15464088651869032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,6144,0.02916711237695482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,2560,0.0063680000603199005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,16384,0.11547999911838108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,1024,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,8192,0.027143999934196472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,2560,0.020202666521072388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,1536,0.0041777777175108595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,768,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,5120,0.013212444053755866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,128,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,7168,0.030151999659008447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,16384,0.045104887750413686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,1536,0.015983111328548856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,32,0.0018195555441909367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,768,0.005978666659858491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,5120,0.01686044368478987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,16384,0.032292445500691734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,3072,0.01274311128589842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,8192,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,256,0.006623111251327727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,7168,0.03067555692460802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,6144,0.027217777238951787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,7168,0.013388444152143268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,3072,0.007373332977294922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,16384,0.0444471107588874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,1536,0.003567111161020067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,3072,0.14102755652533636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,1024,0.009296889106432596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,512,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,3584,0.030525333351559106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,256,0.01684711045689053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,7168,0.10353066523869832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,5120,0.015673778123325773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,2048,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,8192,0.013775111072593264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,768,0.0028391112056043413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,16384,0.07195999887254503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,65536,0.2501342296600342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,2560,0.007028444773621029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,1536,0.00408000002304713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,10240,0.05639644463857015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,768,0.03425333234998915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,65536,0.06388533115386963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,4096,0.007609777980380588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,2048,0.013580444786283704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,512,0.0035013332962989807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,12288,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,6144,0.01921866668595208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,3584,0.032418668270111084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,3584,0.0076453329788313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,10240,0.025422222084469263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,6144,0.041088001595603094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,65536,0.02402488887310028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,16384,0.021369778447681006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,768,0.02084000077512529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,3584,0.00794044468137953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,3584,0.008437333007653555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,128,0.001840888833006223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,64,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,3072,0.008952000074916417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,5120,0.008288000192907121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,2048,0.030730668041441176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,512,0.0042088888585567474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,10240,0.08806311421924168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,7168,0.31333332591586643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,3584,0.010649777948856354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,65536,0.12901688946617976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,16384,0.056074665652381055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,1536,0.0038533334930737815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,128,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,1536,0.018858666221300762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,1024,0.03185866607560052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,10240,0.03137066629197862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,8192,0.021240888370407954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,3072,0.017508443858888414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,6144,0.026533333791626826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,2560,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,6144,0.03427022364404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,6144,0.010197333163685268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,3072,0.011211555865075854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,7168,0.020149333609475028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,2560,0.004957333207130432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,64,0.006312888943486744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,64,0.0021840000732077495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,16384,0.09170399771796332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,768,0.006280888699822956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,512,0.005274666680230035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,8192,0.4265413284301758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,64,0.0025048889219760895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,10240,0.015124445160230001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,16384,0.020071999894248117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,6144,0.024126221736272175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,3584,0.014083556003040738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,8192,0.033639109796947904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,3584,0.008616000413894653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,4096,0.04083199964629279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,768,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,8192,0.01457155578666263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,4096,0.01607911123169793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,256,0.01516711049609714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,16384,0.08001333475112915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,3584,0.014522666732470194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,512,0.10874044232898289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,768,0.009755555954244401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,2560,0.01223111069864697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,3072,0.013802666631009845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,4096,0.004200889004601372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,2048,0.004543111142185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,768,0.005872000008821487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,12288,0.03529689047071669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,4096,0.011517333487669626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,8192,0.052183998955620654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,12288,0.19377422332763672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,3584,0.008239111138714684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,64,0.005910222315125995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,512,0.015471110741297403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,2048,0.008464888566070134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,256,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,5120,0.00925244473748737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,10240,0.027626666757795546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,1536,0.008330666356616551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,128,0.002116444417172008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,2048,0.00656088896923595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,4096,0.014861333701345654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,1024,0.0052168890833854675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,3584,0.010036444498433007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,256,0.01479111115137736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,3072,0.007075555622577667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,64,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,3584,0.022272888157102797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,5120,0.008955555657545725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,1536,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,32,0.01071466671095954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,3072,0.020040889581044514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,128,0.014704000618722705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,1024,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,8192,0.009377777576446533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,4096,0.020569778150982328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,1536,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,3072,0.020078221956888836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,10240,0.004935111022657818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,768,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,2048,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,768,0.02362488872475094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,256,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,64,0.006288888967699475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,128,0.01495733360449473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,16384,0.004994666824738185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,5120,0.01698844465944502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,7168,0.0277804434299469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,4096,0.015156444576051501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,1536,0.005656888915432825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,2048,0.004545777622196409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,128,0.0018435555199782054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,1536,0.019878221882714164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,7168,0.0660239987903171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,6144,0.02737599942419264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,10240,0.014849777022997538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,256,0.035318222310807966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,256,0.014094221923086377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,1536,0.01311466677321328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,2560,0.05001599921120537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,5120,0.024516445067193773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,7168,0.04253510965241326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,2560,0.024738666084077623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,8192,1.7451644473605687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,3072,0.028816888729731243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,10240,1.0060088899400499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,128,0.0347688893477122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,2048,0.007879111501905654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,12288,0.014451555079884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,16384,0.8299235767788358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,1536,0.008987555901209513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,512,0.01808444493346744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,16384,0.08724711338678996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,256,0.08007288641399808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,10240,0.023949333363109167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,6144,0.013328889177905189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,12288,0.019924443629052904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,768,0.019268444842762418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,12288,0.030156443516413372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,2560,0.03272177775700887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,768,0.02178222272131178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,1536,0.06286311149597168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,6144,0.018259555101394653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,128,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,6144,0.009358222285906473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,7168,0.028239998552534316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,7168,0.04159911142455207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,3584,0.09620710876252915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,3072,0.0055440002017551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,10240,0.012422222230169507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,2048,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,6144,0.00872444444232517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,16384,0.03672800130314297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,512,0.019613333874278598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,4096,0.011109333071443768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,256,0.00508622204264005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,2560,0.05221244361665514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,128,0.011349333657158745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,1024,0.003248888792263137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,8192,0.027126222848892212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,4096,0.008635555704434713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,2560,0.00869777798652649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,3584,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,1536,0.005976888868543837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,512,0.012746666868527731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,2048,0.00676711110605134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,768,0.007272889216740926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,512,0.016884444488419425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,1024,0.017862222260899015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,5120,0.02734222180313534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,512,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,256,0.012478222449620565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,1024,0.004852444347408083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,512,0.025555556019147236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,16384,0.01017955525053872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,3584,0.04680266645219591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,5120,0.042275554604000516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,65536,0.22422043482462564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,3584,0.007699555820888943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,3584,0.024655110306209985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,64,0.0017742222795883815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,768,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,3584,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,64,0.006688888702127669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,5120,0.028720888826582167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,16384,0.02605688903066847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,2560,0.014717333846622042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,1536,0.025335111551814612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,128,0.006830222076839871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,256,0.01276711126168569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,512,0.021998221675554912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,128,0.006325333482689328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,1024,0.006025777922736273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,10240,0.015786666009161208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,2048,0.023278221487998962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,10240,0.27095021141899955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,6144,0.013771555489963956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,2560,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,128,0.014458666245142618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,2560,0.024489776955710515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,512,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,2560,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,3072,0.02065333392884996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,3072,0.025594666600227356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,6144,0.023788443870014612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,2560,0.017145777742067974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,8192,0.012404444317022959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,4096,0.0443102220694224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,6144,0.0791964464717441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,768,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,2048,0.019293333093325298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,4096,0.007972444097201029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,4096,0.007796444826655918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,2560,0.007611555357774098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,10240,0.012643555800120035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,768,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,768,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,3584,0.02479644450876448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,8192,0.03914311197068956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,512,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,5120,0.03470933437347412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,64,0.0018195555441909367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,8192,0.03744622071584066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,512,0.0038577777643998465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,65536,0.059338668982187905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,6144,0.015477332803938123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,8192,0.030785779158274334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,128,0.0021271111650599376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,1024,0.00286311118139161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,7168,0.1559386650721232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,256,0.01681066718366411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,16384,0.02113866640461816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,4096,0.008036444584528605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,128,0.2122666703330146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,4096,0.01681600014368693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,4096,0.009109333157539368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,512,0.003536000019974179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,128,0.014470222923490735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,256,0.019304000669055514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,768,0.03173244330618117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,4096,0.0057911111248864075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,512,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,3584,0.004903111192915174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,16384,0.018241778016090393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,1024,0.003565333369705412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,2560,0.027052443888452318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,4096,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,3072,0.01277511070171992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,7168,0.013045333325862885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,4096,0.00796977761718962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,5120,0.08733777867423163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,64,0.005249777601824866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,65536,0.08487021923065186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,8192,0.07275999916924371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,2048,0.018921777606010437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,7168,0.007320000065697565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,256,0.00921688891119427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,1536,0.013779555757840475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,12288,0.03195200032658047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,2048,0.02390755547417535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,16384,0.055867552757263184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,3584,0.013036444783210754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,1536,0.017893332574102614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,2560,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,5120,0.00868800034125646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,65536,0.0716080003314548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,2048,0.09919910960727268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,2048,0.006870222174459034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,1024,0.034113778008355036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,4096,0.014780445231331719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,3072,0.007664889097213745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,3584,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,128,0.015152888165579902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,3072,0.021151999632517498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,16384,0.07111999723646376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,8192,0.03696711195839776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,7168,0.012436444560686747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,768,0.007644444704055786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,3072,0.00996888925631841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,10240,0.27596354484558105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,768,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,10240,0.03458933366669549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,10240,0.022089777721299067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,12288,0.014430221584108142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,32,0.024350222614076402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,768,0.0068488890926043195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,512,0.00286311118139161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,3072,0.018575999471876357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,8192,0.047353777620527476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,10240,0.039954665634367205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,1024,0.009011555876996782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,512,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,1536,0.0038728887836138406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,7168,0.02163822286658817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,768,0.002793777734041214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,3072,0.006973333656787872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,4096,0.01070755554570092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,12288,0.2791671223110623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,256,0.005207111024194294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,1536,0.010029333333174387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,6144,0.023342221975326538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,6144,0.010195555786291758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,3072,0.01959911154376136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,256,0.029382222228580054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,768,0.007295110987292395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,256,0.027386666999922857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,1536,0.005958222266700532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,3072,0.008732444710201686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,10240,0.05836888816621569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,64,0.01033866653839747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,16384,0.039572444227006696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,12288,0.013903111219406128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,65536,0.04684355523851183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,1536,0.04729955395062765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,5120,0.008740444150235917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,16384,0.04292533463901944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,1536,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,1536,0.005143110950787862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,12288,0.013464888764752282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,512,0.0035146665241983202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,7168,0.02958400050799052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,2560,0.007363555332024892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,2560,0.006959111326270633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,2560,0.029382222228580054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,65536,6.200297037760417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,32,0.007675555845101674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,2048,0.13451377550760904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,16384,0.044461333089404635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,5120,0.3183235592312283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,8192,0.011391110718250275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,3584,0.009454222189055549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,1024,0.007276444799370236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,1536,0.01937333411640591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,10240,0.014111111561457316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,1024,0.09530578056971233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,6144,0.12721778286827937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,3072,0.010312889185216691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,5120,0.010792888700962067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,128,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,512,0.0034906665484110513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,65536,0.058768888314565025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,128,0.005199111170238919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,2048,0.00666844430896971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,1536,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,8192,0.011046222514576383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,5120,0.018837332725524902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,8192,0.017867555220921833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,256,0.021408889028761122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,16384,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,7168,0.01680444512102339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,1536,0.01308444473478529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,10240,0.05868888563579983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,1536,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,128,0.014506666196717156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,7168,0.11308266056908502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,32,0.0018595556418100994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,256,0.005964444329341252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,256,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,128,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,512,0.0031635556370019913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,64,0.002534222271707323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,3072,0.015079110860824585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,5120,0.08916799889670478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,64,0.002151111140847206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,256,0.025253333979182776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,512,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,1536,0.008354666332403818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,65536,0.06952622201707628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,128,0.0069146665434042616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,7168,0.09882577922609116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,6144,0.009766221874290043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,6144,0.018775110443433125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,2560,0.01147466649611791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,1024,0.0035235554807715942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,32,0.0021422221842739317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,12288,0.019295111298561096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,2048,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,64,0.00281866660548581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,4096,0.011009777585665384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,768,0.018933332628673978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,2560,0.02608799934387207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,1536,0.02387822171052297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,128,0.01072088877360026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,16384,0.4091493288675944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,256,0.004872888740566042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,8192,0.010689777632554373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,1536,0.0069715554515520734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,256,0.003565333369705412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,128,0.011744000017642975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,128,0.013472889032628802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,1024,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,3072,0.008403555386596257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,512,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,2048,0.009454222189055549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,7168,0.012039999994966718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,3584,0.01237422227859497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,4096,0.008310221963458592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,3584,0.011032889286677042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,4096,0.00979288915793101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,5120,0.010701333483060202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,2560,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,65536,0.14628088474273682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,4096,0.016154666741689045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,10240,0.04459644357363383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,12288,0.01684266659948561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,2048,0.027297778262032404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,512,0.0028204443968004654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,2048,0.014102222190962898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,2560,0.020735111501481798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,512,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,12288,0.03616711166169908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,12288,0.05570399761199951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,16384,0.028536889288160536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,12288,0.06038933330112033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,3584,0.011295999917719098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,5120,0.020290666156344943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,5120,0.025058666865030926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,512,0.007085333267847697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,12288,0.014495111174053617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,65536,0.09091733561621772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,1024,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,5120,0.013061333033773633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,2048,0.007181333170996771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,1024,0.02331200076474084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,128,0.05877955754597982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,512,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,16384,0.016132444143295288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,768,0.015842666228612263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,1024,0.01515111161602868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,6144,0.03180088930659824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,2560,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,3072,0.004551110996140374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,2048,0.007322666545708974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,8192,3.0022941165500217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,4096,0.04193066557248434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,7168,0.07169155279795329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,8192,0.013659555878904132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,2560,0.009296000003814697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,10240,0.008987555901209513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,16384,6.3143403795030375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,1536,0.0063528890411059065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,10240,0.02794577678044637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,512,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,12288,0.014763555592960782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,16384,0.03153333399030898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,5120,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,3072,0.00756444450881746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,3072,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,2560,0.03236177894804213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,256,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,3584,0.0239964442120658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,12288,0.06053421894709269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,12288,0.021740444832377966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,1024,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,128,0.005277333160241445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,512,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,1024,0.0188773340649075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,65536,0.01647200021478865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,65536,0.22346756193372938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,1536,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,2048,0.059112886587778725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,12288,0.40507645077175564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,256,0.02552977701028188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,7168,0.10660088724560207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,3072,0.0073902226156658595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,65536,0.1841519938574897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,7168,0.028186665640936956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,1024,0.01311822235584259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,3072,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,512,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,7168,0.014508444401952954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,128,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,128,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,6144,0.01854311095343696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,3584,0.007786666353543599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,3072,0.007686222592989604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,6144,0.04152889053026835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,12288,0.016474665866957772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,65536,0.21161866188049316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,12288,0.04408266809251574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,8192,0.024713777833514746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,65536,0.06242844131257799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,8192,0.02438755498992072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,3072,0.01000711073478063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,8192,0.004916444420814514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,32,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,768,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,128,0.005280888742870755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,3584,0.022084444761276245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,8192,0.014564444621404013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,128,0.015106666419241162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,64,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,512,0.005199111170238919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,1536,0.020276443825827706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,3584,0.03397422366672092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,128,0.014826666977670459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,10240,0.02129688858985901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,3072,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,3072,0.026144888665941026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,2048,0.1144675546222263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,1024,0.006220444622966979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,128,0.0019991110182470745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,512,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,512,0.006050666587220297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,3072,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,256,0.016504888733228047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,4096,0.034456888834635414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,1536,0.01683111157682207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,512,0.005203555441564984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,256,0.006562666760550604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,65536,0.13391378190782335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,1536,0.004969777746333016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,3072,0.01772177716096242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,5120,0.008588444855478074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,128,0.006464889066086874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,65536,0.10154933399624294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,10240,0.04870311088032193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,3584,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,8192,0.030547555949952867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,6144,0.01072622256146537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,256,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,3584,0.010488889283604093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,256,0.010701333483060202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,128,0.0018791111393107309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,1536,0.013842666314707862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,2048,0.016262221667501662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,2560,0.008995555341243744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,2048,0.0535662207338545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,1024,0.005816000203291575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,128,0.014158222410413953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,3072,0.025739555557568867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,64,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,1024,0.015853333804342482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,5120,0.009303111169073317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,5120,0.12866222858428955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,8192,0.022679999470710754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,768,0.003519111209445529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,10240,0.04959288901752896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,32,0.004534222185611725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,8192,0.012903999951150684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,16384,0.0346959994898902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,1024,0.020233778489960563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,8192,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,7168,0.07451733615663317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,3072,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,65536,0.07353778017891778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,2560,0.007307555940416124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,12288,0.010699555277824402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,5120,0.06240799691942003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,12288,0.8234160211351184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,2560,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,12288,0.04526577724350823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,768,0.0069146665434042616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,1536,0.018929777873886954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,32,0.004622222234805425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,2048,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,6144,0.00628977765639623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,5120,0.022954665952258643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,16384,0.04461866617202759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,512,0.006018666757477655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,128,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,3584,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,6144,0.009201777478059133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,12288,0.03908088803291321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,4096,0.009005332986513773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,1024,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,32,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,768,0.005953777581453323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,7168,0.035785777701271906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,2560,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,6144,0.21249332692888048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,2560,0.029537777105967205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,256,0.00999111102686988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,4096,0.03575377662976583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,10240,0.8151875601874458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,512,0.007965332931942409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,3072,0.021012443635198805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,2048,0.015460444821251763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,6144,0.1279333300060696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,7168,0.03424266642994351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,1024,0.018223999275101554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,10240,0.03368799885114034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,4096,0.016131555040677387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,3584,0.010043555663691627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,1024,0.03677777780426873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,4096,0.08829244640138413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,5120,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,64,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,512,0.016535111599498324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,256,0.0052168890833854675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,512,0.022942221826977197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,3072,0.02813777658674452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,32,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,8192,0.02516088883082072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,16384,0.023955555425749883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,6144,0.004592000196377437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,768,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,2560,0.007634667058785756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,128,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,4096,0.015853333804342482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,1536,0.011049778097205691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,6144,0.017824000782436795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,10240,0.0540924436516232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,8192,0.010904889139864178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,1536,0.008639111287064022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,768,0.004288889053795072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,65536,0.20976621574825713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,6144,0.02906399965286255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,10240,0.029249777396519978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,3072,0.01684533390733931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,768,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,5120,0.01718755563100179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,65536,0.1346097787221273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,1024,0.028528001573350694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,5120,0.2545128928290473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,2048,0.02436888880199856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,512,0.010030222435792288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,4096,0.008434666527642144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,5120,0.008952000074916417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,2560,0.008256889051861232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,128,0.005208000126812193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,768,0.017691555950376723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,12288,0.028590222199757893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,768,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,512,0.0031866667171319327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,65536,0.4447493553161621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,4096,0.12025422520107693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,512,0.005630222045713001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,2048,0.008990222381220924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,3072,0.023200889428456623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,1024,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,8192,0.014208000567224292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,32,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,128,0.009931555224789513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,256,0.0021786666992637846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,12288,0.014491554763582019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,2560,0.005669333454635408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,10240,0.020609777834680345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,65536,0.01683111157682207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,3072,0.008972444468074376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,3072,0.028464890188641016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,2560,0.007048000064161088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,128,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,3584,0.007247111035717859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,16384,0.027403554982609216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,3072,0.013173333472675748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,10240,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,128,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,3072,0.017849778135617573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,16384,0.016182222300105624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,5120,0.10397511058383518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,10240,0.03227377931276957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,7168,0.020390222469965618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,3584,0.009621333744790819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,64,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,10240,0.048935999472936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,3584,0.006903111106819577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,6144,0.019167110323905945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,2048,0.004914666629499859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,7168,0.015032889114485847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,1024,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,768,0.01237422227859497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,8192,0.03124622172779507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,512,0.0041759999261962045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,768,0.00849422232972251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,8192,0.024833778540293377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,768,0.0028595555987623003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,6144,0.05871733029683431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,256,0.0021679999513758556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,2560,0.007742221984598372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,3584,0.007511110769377814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,64,0.0021253333737452826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,65536,0.06256177690294054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,32,0.0018355554590622585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,12288,0.01685333251953125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,3072,0.007242666350470648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,12288,0.02089866664674547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,4096,0.00999466660949919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,768,0.013824000126785703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,5120,0.014380445082982382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,12288,0.04565422071350945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,4096,0.018146667215559218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,3072,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,4096,0.014150222142537435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,512,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,512,0.018187556001875136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,512,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,1536,0.0063031112982167145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,32,0.0018168888572189545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,8192,0.014049778381983439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,5120,0.011244444383515252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,128,0.005390222287840313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,512,0.002463999928699599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,256,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,3584,0.010671999719407825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,4096,0.01923022170861562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,64,0.0018906665758954154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,512,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,3072,0.015573332707087198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,3584,0.009983110758993361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,5120,0.03014311194419861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,3072,0.011239999698268043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,6144,0.011540444360838996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,65536,0.31350665622287327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,1536,0.018224888377719455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,16384,0.17457511689927843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,1024,0.006030222194062338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,256,0.00309955556359556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,65536,0.13334577613406712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,1024,0.18458933300442168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,6144,0.007956444389290279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,7168,0.1373013390435113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,128,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,7168,0.01036177741156684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,7168,0.020261334048377145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,65536,0.17906667126549614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,6144,0.009688888986905416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,2048,0.46440887451171875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,16384,0.040539556079440646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,65536,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,5120,0.01719111038578881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,1024,0.016912889149453905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,12288,0.005208000126812193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,768,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,512,0.002522666628162066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,6144,0.1296959982977973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,16384,0.07225777705510457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,65536,0.06724266873465644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,1024,0.0069022224181228215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,512,0.002480000050531493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,3072,0.08912711011038886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,1024,0.007979555262459649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,1024,0.006605333338181178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,1024,0.016515556308958266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,8192,0.014037332601017423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,16384,0.5677013397216797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,2048,0.02373955481582218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,768,0.015343111422326831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,4096,0.0069022224181228215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,3584,0.005698666804366642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,4096,0.007977777885066139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,128,0.005472888963090048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,10240,0.034119112624062434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,128,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,1024,0.018907555275493197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,4096,0.014905777242448596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,4096,0.14400622579786512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,7168,0.021249777740902368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,128,0.005048000150256687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,65536,0.07349955373340182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,6144,0.3249928951263428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,2560,0.006590222318967183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,7168,0.01275644451379776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,3584,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,3072,0.007322666545708974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,3072,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,768,0.007632888853549957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,2560,0.025034666061401367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,65536,1.3714817894829645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,6144,0.0488391121228536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,4096,0.03363200028737386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,7168,0.07172089152865939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,2048,0.05497689048449198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,512,0.013404444687896304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,32,0.005490666462315454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,8192,0.23037865426805285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,5120,0.031180444690916274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,1024,0.012919999659061432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,5120,0.015474667151769003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,10240,0.02581866747803158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,10240,0.021378666162490845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,16384,0.04611911045180427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,10240,0.5263937844170464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,1024,0.0041164445380369825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,12288,0.07531822390026517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,2048,0.023588443795839947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,2560,0.015491555134455362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,2048,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,512,0.009025777379671732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,1024,0.014169777433077494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,16384,0.04249777727656894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,2048,0.01443288889196184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,6144,0.012120889292822944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,10240,0.010370666782061258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,2560,0.0343537794219123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,1024,0.0057502223385704895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,10240,0.10089866982565986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,1024,0.01853155593077342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,4096,0.008436444732877944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,128,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,2560,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,4096,0.02661955522166358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,16384,0.20245867305331758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,10240,0.031529777579837374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,3584,0.00869688888390859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,2048,0.004592888885074192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,6144,0.006825777805513806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,512,0.004556444370084339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,512,0.007629333270920648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,5120,0.012660444610648684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,6144,0.036679112248950534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,768,0.013087111214796701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,128,0.015459555718633862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,64,0.001812444378932317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,256,0.0024968888610601425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,8192,0.015767110718621146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,32,0.0057608890864584185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,64,0.08911110957463582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,4096,0.009659555223253038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,16384,0.0926213330692715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,10240,0.0984613365597195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,5120,0.032159109910329185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,32,0.0020391111158662373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,16384,0.021889777647124395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,16384,0.024029332730505202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,10240,0.008683555656009251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,3584,0.24994399812486437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,32,0.007323555648326874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,5120,0.008967111508051554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,1024,0.010595555934641095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,65536,0.1213466723759969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,2560,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,7168,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,4096,0.00831200016869439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,10240,0.10062133603625828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,768,0.03276977936426798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,256,0.006593777901596493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,4096,0.06994577911165026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,8192,0.018882667024930317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,12288,0.01686311099264357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,4096,0.018186666899257235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,10240,0.012715555727481842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,1536,0.0038657776183552216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,6144,0.16428800423940024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,6144,0.009302222066455418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,65536,0.12280000580681695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,1536,0.08478755421108669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,7168,0.015027556154463025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,3584,0.19530310895707872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,7168,0.038380444049835205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,3584,0.007890666524569193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,1536,0.18102667066786024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,512,0.018915555543369718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,4096,0.016389333539538913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,2048,0.0074133334888352295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,768,0.007751999629868402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,16384,0.03152444296412998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,12288,0.04235733217663235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,65536,0.06129599942101372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,8192,0.016494222813182406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,16384,0.2892213397555881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,12288,0.01650044487582313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,6144,0.08909066518147786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,32,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,12288,0.04968622326850891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,3584,0.10272622108459473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,4096,0.02496355606449975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,65536,0.13457689020368788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,3072,0.015823110938072205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,32,0.0017662222186724346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,768,0.006226666685607698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,7168,3.2235537634955516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,256,0.002867555452717675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,256,0.02057066725360023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,512,0.01645955608950721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,4096,0.02880711025661892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,16384,0.15051289399464926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,5120,0.038239111502965294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,7168,0.03328266739845276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,1024,0.020943111843532983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,10240,0.010772444307804108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,65536,0.06118933359781901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,16384,0.038915554682413735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,16384,0.019695111446910434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,2560,0.02630222174856398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,2048,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,512,0.0151235560576121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,3584,0.004912888838185204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,768,0.03608355588383145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,512,0.005274666680230035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,6144,0.01905866629547543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,2048,0.06214400132497152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,64,0.0028231110837724474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,12288,0.019955555597941082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,5120,0.038344889879226685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,10240,0.05355911122428047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,65536,0.09080000056160821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,1024,0.005958222266700532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,768,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,256,0.0052764444715446895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,512,0.012042666474978128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,3072,0.004530666602982415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,4096,0.0079297779334916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,256,0.05918222003512912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,3072,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,3584,0.008299555215570662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,10240,0.012042666474978128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,768,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,12288,0.05970578061209785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,5120,0.008678221868144141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,512,0.002513777878549364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,6144,0.009294222626421187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,4096,0.012136889000733694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,7168,0.023687111006842718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,2048,0.004182222402758068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,768,0.004971555537647671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,10240,0.012416889270146688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,12288,0.013796444568369122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,8192,0.12026577525668675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,1536,0.008355555435021719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,3072,0.0528417792585161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,2048,0.006863111009200414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,3072,0.013864888913101621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,256,0.04493955439991421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,128,0.0088746663596895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,768,0.0031893334041039148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,2560,0.006027555714050929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,1024,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,64,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,12288,0.019923556182119582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,2048,0.038359112209743924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,8192,0.013680000272062091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,8192,0.023507555325826008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,32,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,128,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,512,0.012800888882742988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,1024,0.00314666661951277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,1024,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,3072,0.010019555687904358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,16384,0.016217778126398723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,10240,0.0671280026435852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,4096,0.035214222139782377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,256,0.01516533394654592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,4096,0.008639111287064022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,1024,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,32,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,65536,0.14578933186001247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,3584,0.007308444215191736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,16384,1.1388764911227756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,65536,0.0912053320142958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,16384,0.21895910633934867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,8192,0.012318222059143914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,3072,0.010048888623714447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,768,0.009112888740168678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,512,0.00793422261873881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,1024,0.011693333586057028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,3584,0.00793599999613232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,256,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,768,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,3072,0.02846577763557434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,4096,0.026455110973782007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,3072,0.007428444094128079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,1536,0.0064488889442549805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,5120,0.008991999758614434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,16384,0.017137777474191453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,2560,0.0997919970088535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,12288,0.028455999162462022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,768,0.008676444490750631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,128,0.004442666553788715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,6144,0.039389332135518394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,2560,0.025566221939192876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,4096,0.01605333387851715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,768,0.0038328886859946777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,512,0.009124444590674506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,128,0.10877422491709392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,65536,0.12467643949720596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,5120,0.015896000795894198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,8192,0.017137777474191453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,12288,0.05109511150254143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,1536,0.011720000041855706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,3072,0.00997866690158844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,2048,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,6144,0.07295999924341838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,12288,0.026072000463803608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,3072,0.007813333637184566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,7168,0.053132444620132446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,6144,0.15608178244696722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,12288,0.01369511087735494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,4096,0.008389333056079017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,5120,0.010783111055692038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,2560,0.01998666591114468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,512,0.0031439999325407874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,1536,0.008276444342401292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,512,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,7168,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,512,0.03336266676584879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,256,0.008640888664457533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,5120,0.00869422240389718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,64,0.0018648888087934917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,1024,0.0028284444577164124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,65536,0.0698195563422309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,1024,0.007253333098358578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,65536,1.251106686062283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,7168,0.014151111245155334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,65536,0.04075466593106588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,8192,0.06196710798475477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,8192,0.14423378308614096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,8192,0.013680000272062091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,2560,0.007191999918884701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,2560,0.006959111326270633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,256,0.007992888490358988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,5120,0.008950221869680617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,3072,0.025581333372328017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,8192,0.014577777849303352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,3584,0.004935111022657818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,7168,0.08024888568454318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,16384,0.021180444293551978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,16384,0.20844354894426134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,3584,0.03698577814631992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,768,0.00254488881263468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,1024,0.005817777580685086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,16384,0.03474044468667772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,1536,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,3584,0.08278755346934001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,1024,0.019564444820086162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,512,0.006335999816656113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,8192,0.010715555813577441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,3072,0.01089333328935835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,64,0.006887110984987683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,7168,0.043288889858457774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,6144,0.01964000033007728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,1536,0.018534221582942538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,3584,0.014800000521871777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,7168,0.020466667082574632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,2048,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,4096,0.009542221824328104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,256,0.013413333230548434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,7168,0.021345777644051447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,1536,0.05951199928919474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,4096,0.040607998768488564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,5120,0.008341333104504479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,3072,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,3584,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,3072,0.017536888519922893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,10240,0.0177208897140291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,2560,0.051388445827696055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,128,0.014749333262443542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,1024,0.0028640000770489373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,16384,0.043084445926878184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,1536,0.01124888906876246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,16384,0.028804444604449805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,256,0.005190222213665645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,6144,0.14769955476125082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,6144,0.008963555925422246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,2560,0.022464889619085524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,7168,0.029842668109469946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,12288,0.0821031133333842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,3072,0.01311555587583118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,10240,0.012376888758606382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,2048,0.03350844317012363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,768,0.01991911066903008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,768,0.014483556151390076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,1024,0.01629688839117686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,128,0.013456888496875763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,3072,0.013390222357379066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,4096,0.020928889513015747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,256,0.016173332929611206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,16384,0.7418026924133301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,7168,0.036357333262761436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,10240,0.126127110587226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,4096,0.01127822200457255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,7168,0.015132443772421943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,2048,0.0062977779242727495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,10240,0.05924533473120796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,3584,0.016128000285890367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,3072,0.015457777513398064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,3072,0.014575110541449653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,512,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,6144,0.13167288568284777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,2560,0.020503999458418954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,128,0.003216000066863166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,5120,0.01216533366176817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,12288,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,1024,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,768,0.012808889150619507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,6144,0.010341333018408881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,2048,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,3584,0.010802666346232096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,65536,0.29679377873738605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,12288,0.013455111119482251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,128,0.01072177787621816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,16384,0.23878577020433214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,16384,0.3398426638709174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,8192,0.02353688908947839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,6144,0.01752088963985443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,32,0.0018568889548381169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,10240,0.020255110330051847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,2048,0.021033777130974665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,3584,0.01683644453684489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,1024,0.0041964443193541635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,12288,0.0336497790283627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,5120,0.009405333134863112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,32,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,65536,0.18388266033596465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,2048,0.005364444520738389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,7168,0.019573332534896005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,7168,0.010235555469989777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,1536,0.006513777706358168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,10240,0.03979466689957513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,768,0.017528888252046373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,5120,0.01789155602455139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,256,0.005971555494599872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,1536,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,8192,0.010764444039927589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,3072,0.013420444395807056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,4096,0.10036622153388129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,4096,0.013311111264758639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,1024,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,7168,0.020662221643659804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,2048,0.013752889302041797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,2048,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,10240,0.03527466787232293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,5120,0.02144266664981842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,2048,0.052170667383405894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,512,0.018556444181336295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,512,0.01607911123169793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,4096,0.28528979089525014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,16384,0.07854133182101779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,10240,0.013759111364682516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,6144,0.31775376531812877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,4096,0.024819556209776137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,6144,0.00903555585278405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,512,0.017892445127169292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,3584,0.01500266624821557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,1536,0.032277332411872015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,4096,0.01812266641192966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,65536,0.19177688492668998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,16384,0.02939200070169237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,64,0.0052897776994440295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,65536,0.9121066199408637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,10240,0.12658133771684435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,2560,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,16384,0.04042488998836941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,512,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,2048,0.010010666317409938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,10240,0.02613688839806451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,2560,0.015456888410780164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,8192,0.10617866781022813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,768,0.024127110838890076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,65536,0.3057191107008192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,8192,0.024567110670937434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,7168,0.17835377322302925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,6144,0.014123555686738757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,3584,0.02051377793153127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,3584,0.34775198830498594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,4096,0.007815111014578078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,4096,0.010664888554149203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,8192,0.016295111841625638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,16384,0.02141688929663764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,256,0.01685688893000285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,8192,0.023288889063729182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,6144,0.009022222624884712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,32,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,2560,0.004875555634498596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,64,0.0017831110292010838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,32,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,16384,0.01607288916905721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,65536,0.027385777897304956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,6144,0.1715093321270413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,32,0.0022133334229389825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,1536,0.003583999971548716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,5120,0.33527112007141113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,2048,0.014627556006113688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,16384,0.01178311142656538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,3072,0.013528000149461957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,7168,0.01998044384850396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,4096,0.015800888339678448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,3584,0.028704888290829126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,16384,0.018794667389657762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,8192,0.014628445108731588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,5120,0.02089244458410475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,10240,0.21085333824157715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,3584,0.007301333049933116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,65536,0.05211288730303446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,512,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,256,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,128,0.04633422361479866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,2048,0.03148355417781406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,3072,0.011352889239788055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,512,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,5120,0.021585777401924133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,6144,0.024544888072543677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,1024,0.0032097777972618737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,512,0.031698667340808444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,65536,0.3107271194458008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,2048,0.010943111446168689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,16384,0.02350933353106181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,12288,0.02809244394302368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,4096,0.11790843804677327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,4096,0.004903111192915174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,8192,0.038823998636669584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,5120,0.009319999979601966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,5120,0.12061422401004368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,65536,0.06649866369035509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,512,0.0028320000403457214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,32,0.003160888950030009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,8192,0.01993688941001892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,4096,0.01585422290696038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,2048,0.028415110376146104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,1536,0.008301333420806462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,65536,1.0289395650227864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,6144,0.023015111684799194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,4096,0.027056889401541814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,2048,0.018255111243989732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,256,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,512,0.002473777780930201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,2048,0.011628444823953839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,1536,0.006907555378145642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,4096,0.009337777892748514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,16384,0.014648889501889547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,6144,0.011209777659840055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,128,0.014463111758232117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,768,0.09259555737177531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,12288,0.19551200336880156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,2048,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,8192,0.010682666467295753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,16384,0.02148799929353926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,65536,0.16082666979895696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,3584,0.018783999813927543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,5120,0.025991110338105097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,5120,0.016944888565275405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,3584,0.02283199959331089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,5120,0.1475386619567871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,1536,0.01891288823551602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,8192,0.014187556174066333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,64,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,1536,0.011383111278216044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,8192,0.01475822263293796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,2048,0.016175111134847004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,768,0.010858666565683154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,128,0.005038222091065513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,3584,0.008309333688682979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,8192,0.010674667027261523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,128,0.0052471111218134565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,128,0.0018159999615616268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,256,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,1024,0.013106666505336761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,6144,0.006294222341643439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,2560,0.004920888692140579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,3072,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,1536,0.01585244470172458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,512,0.023237332701683044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,16384,0.029840889904234145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,5120,0.016399110356966656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,7168,0.2925288942125109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,6144,0.07375644312964545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,10240,0.029134223858515423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,1536,0.008585777547624376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,2560,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,3584,0.005176888985766305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,512,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,7168,0.046105778879589505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,6144,0.12995644410451254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,768,0.021280889709790547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,768,0.016568889220555622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,65536,0.06619111034605238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,768,0.007582222421964009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,16384,0.21169422732459175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,512,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,512,0.08463022443983291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,5120,0.01205599970287747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,5120,0.010075555907355415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,128,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,7168,0.019138667318556044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,7168,0.014923555983437432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,3584,0.004905777672926585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,64,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,6144,0.01440533333354526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,32,0.0014728889283206726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,3584,0.021141333712471858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,768,0.004520888957712385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,512,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,512,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,7168,0.010493333141009012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,64,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,16384,0.01958755486541324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,4096,0.007702222300900354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,4096,0.007928888830873702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,7168,0.017489777670966256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,16384,0.06615466541714139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,5120,0.009686222506894005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,7168,0.024007999234729346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,65536,0.06203733550177681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,3072,0.00794311116139094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,128,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,65536,0.07139466868506537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,16384,0.011019555230935415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,3584,0.026386666629049513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,256,0.004907555464241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,2048,0.004942222187916438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,512,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,768,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,65536,0.25594666269090444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,2048,0.006817777951558431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,768,0.01312888910373052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,8192,0.014559111661381192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,1536,0.0037662221325768363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,3584,0.011253332926167382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,2048,0.01991911066903008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,6144,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,16384,0.2880035506354438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,3584,0.025014221668243408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,768,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,768,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,8192,0.012090666426552666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,1024,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,12288,0.027122666438420612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,8192,0.010962666736708747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,12288,0.03224888775083754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,10240,0.011939555406570435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,3584,0.010261333651012844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,256,0.0025155554629034465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,3072,0.059060447745853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,16384,0.018611555298169453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,512,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,16384,0.05674755573272705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,768,0.005613333235184352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,2048,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,3584,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,2560,0.007505777809354994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,2048,0.03465155429310269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,3072,0.008597333398130205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,65536,0.1513813336690267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,32,0.0120666664507654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,8192,0.013775111072593264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,65536,0.07782489061355591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,768,0.02163911031352149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,8192,0.005420444326268301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,64,0.017771555317772757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,512,0.0031662223239739737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,8192,0.023729777998394434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,8192,0.050372444921069674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,4096,0.00794044468137953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,128,0.050211555427975126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,768,0.02495199938615163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,65536,0.12178133593665229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,65536,0.018568888306617737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,2560,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,5120,0.14517154958513048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,5120,0.009300444689061906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,6144,0.026662222213215295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,3584,0.03393422232733832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,6144,0.018778666853904724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,768,0.019932443896929424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,65536,0.062498668829600014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,10240,0.012808889150619507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,512,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,3072,0.010232888989978367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,64,0.010909332997269101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,6144,0.03053599927160475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,4096,0.02372444503837162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,3072,0.03702488872740004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,16384,0.020563556088341605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,16384,0.07690044244130452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,8192,0.010464888479974536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,5120,0.016759110821617972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,256,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,32,0.0018275556051068837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,1536,0.009353777600659264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,4096,0.008228444390826756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,32,0.0021840000732077495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,16384,0.13432621955871582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,2048,0.03016622198952569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,3584,0.017551110850440133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,5120,0.015788444214397006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,5120,0.008105778031879002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,4096,0.004520888957712385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,5120,0.009024000002278222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,3072,0.007585778004593319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,16384,0.16184533966912165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,3072,0.007786666353543599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,5120,0.0111191107167138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,3072,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,12288,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,16384,0.05020622081226773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,3584,0.014979556202888489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,1024,0.003527111063400904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,10240,0.013058666553762225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,16384,0.05297599898444282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,65536,0.3575982252756755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,256,0.008009778128729926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,3072,0.16200978226131862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,2048,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,2048,0.358096890979343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,256,0.016869333055284288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,3072,0.009020444419648912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,512,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,3072,0.01718133356836107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,7168,0.07996800210740831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,7168,0.014764444695578681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,2048,0.04753511150677999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,6144,0.15051289399464926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,3584,0.043619556559456714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,512,0.015152888165579902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,3584,0.2785173257191976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,768,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,2560,0.00981866651111179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,10240,0.039847112364239164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,16384,0.03025955624050564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,3584,0.01578044394652049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,8192,0.015043555034531487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,7168,0.017887999614079792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,256,0.006224888894293044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,65536,0.03151822090148926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,3072,0.015325332681337992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,7168,0.24767555130852592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,256,0.0021235555824306277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,10240,0.01310755560795466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,128,0.010993777877754636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,3072,0.0075777777367168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,128,0.005578666511509154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,2560,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,256,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,32,0.0018355554590622585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,65536,0.2298355632358127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,7168,0.07748266723420885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,4096,0.01481244464715322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,7168,0.014103111293580798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,2048,0.07950933112038507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,8192,0.019623999794324238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,12288,0.019927110936906602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,3584,0.07395733065075345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,6144,0.00960355583164427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,3584,0.007507555186748505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,16384,0.016963554753197562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,5120,0.008540444076061249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,512,0.0058666666348775225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,12288,0.017862222260899015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,7168,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,32,0.0018044445249769424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,16384,0.036554667684766985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,512,0.0068915556702348925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,1024,0.002875555513633622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,3584,0.018167111608717177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,3072,0.052829331821865506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,65536,0.5284719996982151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,2560,0.009469333622190688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,12288,0.020404444800482858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,64,0.00693866651919153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,32,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,768,0.01442311041884952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,1536,0.00775999989774492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,16384,0.5908257696363661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,16384,0.1042426692114936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,128,0.015096889601813423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,8192,0.02741333345572154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,12288,0.030328889687856037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,10240,0.053167111343807645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,1024,0.003847111016511917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,3072,0.0811742213037279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,7168,0.012119111087587146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,7168,0.14163733853234187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,10240,0.016632889707883198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,3072,0.009532444179058075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,32,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,6144,0.015243555108706156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,768,0.00286311118139161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,128,0.005642666584915585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,1024,0.015133332875039844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,1024,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,16384,0.049201776583989464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,2560,0.00490488898422983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,128,0.010131555298964182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,1536,0.016199999385409884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,256,0.016155555844306946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,256,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,6144,0.04107466671201918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,6144,0.06933066580030653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,65536,0.09047910902235244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,1024,0.004204444587230682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,3072,0.05311110946867201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,768,0.007639110916190677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,7168,0.009688888986905416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,768,0.023775110642115276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,6144,0.10072800185945298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,65536,0.0563031103875902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,2560,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,4096,0.02341777748531765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,1536,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,128,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,512,0.01991911066903008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,768,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,7168,0.0396106673611535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,512,0.027497778336207073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,3584,0.009663999908500248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,128,0.005493333356248007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,256,0.012734221915404001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,65536,0.06749955813090007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,256,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,256,0.0022408889813555614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,10240,0.025817778375413682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,12288,0.12402666939629449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,12288,0.15236800246768528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,65536,0.020256888535287645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,512,0.005954666684071223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,8192,0.027776888675159876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,1024,0.06789244545830621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,10240,0.3199937873416477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,4096,0.007888000044557784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,512,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,3072,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,12288,0.03478577733039856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,16384,0.22975020938449434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,16384,0.01869155632125007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,8192,0.015091554986106025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,8192,0.07942489120695326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,256,0.0031911111954185697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,7168,0.016223111086421542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,768,0.01871288816134135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,65536,0.24724356333414713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,6144,0.015847999188635085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,8192,0.016328000360065036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,65536,0.09259288840823704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,1536,0.06855822271770902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,8192,0.0347688893477122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,1536,0.26528088251749676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,128,0.0021688888470331826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,128,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,16384,0.039761778381135725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,2048,0.009995555712117089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,12288,0.022315555148654517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,3584,0.007624000310897827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,16384,0.06430221928490533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,10240,0.023639111055268183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,16384,0.2764666610293918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,256,0.002124444478087955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,3072,0.01407999959256914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,1536,0.015126221709781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,5120,0.029288000530666772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,1536,0.0032071111102898917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,128,0.0021928888228204516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,64,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,3584,0.007696888513035244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,2560,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,2560,0.013439111411571503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,10240,0.020799110333124798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,65536,1.9481563568115234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,5120,0.012054222325483957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,32,0.009346666435400644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,12288,0.014426666829321118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,1536,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,4096,0.09688266780641343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,10240,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,64,0.001793777777089013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,1536,0.0041697778635554844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,768,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,768,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,2048,0.004876444323195351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,256,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,5120,0.021809778279728357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,6144,0.10944888989130656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,4096,0.009287111461162567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,256,0.006565333240562015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,4096,0.08390133248435126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,5120,0.04316533274120755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,12288,0.012086222569147745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,3072,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,2560,0.004914666629499859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,256,0.0021955555097924336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,12288,0.14914845095740423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,7168,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,1024,0.008275555239783393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,1024,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,128,0.006919999917348226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,32,0.01720888912677765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,2560,0.007320889168315464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,2048,0.014815110299322339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,768,0.07398666938145955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,65536,0.05805689096450806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,6144,0.024777778320842322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,1024,0.006223111102978389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,16384,0.040101332796944514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,8192,0.009714666340086196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,512,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,4096,0.015824000040690105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,768,0.014022222823566861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,8192,0.0911102228694492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,6144,0.02146400014559428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,2560,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,3584,0.008605333666006723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,4096,0.01662133302953508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,1024,0.0058755555914507965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,64,0.002453333387772242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,256,0.00249777775671747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,512,0.21673066086239287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,65536,0.7412835756937662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,2048,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,1024,0.03562755717171563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,4096,0.014313777287801107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,512,0.006728000111050076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,16384,0.02401688860522376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,5120,0.01940622263484531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,16384,0.06005422274271647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,5120,0.012516444755925072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,12288,0.14971200625101724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,6144,0.004916444420814514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,3584,0.016737777325842116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,128,0.0028364445186323593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,512,0.005362666729423735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,5120,0.01664533383316464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,768,0.006212444355090459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,3072,0.07405866516960992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,2560,0.0329422222243415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,5120,0.018545778261290658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,768,0.01750666730933719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,2560,0.008567999634477828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,128,0.015453333655993143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,8192,0.012630222572220696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,4096,0.040377779139412776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,128,0.015248888068728976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,3072,0.011420444481902652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,16384,0.08232000139024523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,6144,0.025039111574490864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,16384,0.05871911181343926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,3072,0.008547555241319869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,8192,0.3250533209906684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,128,0.011036444041464063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,16384,0.02106222179200914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,512,0.003167111012670729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,512,0.002577777744995223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,10240,0.035946667194366455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,3584,0.0114720000161065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,7168,0.012930666406949362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,4096,0.009658666948477427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,1536,0.0768595536549886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,256,0.002469333302643564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,7168,0.01811911165714264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,5120,0.05764000283347236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,16384,0.02379111117786831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,768,0.022593778040673997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,3072,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,65536,0.07652444309658475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,6144,0.021213332811991375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,768,0.006635555376609166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,10240,0.010047111246320937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,7168,0.02312622302108341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,10240,0.11280088954501682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,3072,0.008578666382365758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,5120,0.008976889153321585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,3584,0.017477333545684814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,16384,0.027217777238951787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,3584,0.0058782220714622075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,128,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,4096,0.016569778323173523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,128,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,3584,0.013998222019937305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,128,0.01511111193233066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,7168,0.06604799959394667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,128,0.005672889037264719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,3072,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,7168,0.02938400043381585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,1024,0.01756000022093455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,6144,0.005208888815508948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,65536,0.12918133205837673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,1536,0.006804444309737947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,3584,0.05719644493526883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,1024,0.025633777181307476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,12288,0.01733688844574822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,10240,0.050851556989881724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,2048,0.0185217774576611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,4096,0.17941066953870985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,128,0.020000888241661918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,5120,0.01993777851263682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,1024,0.03600533472167121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,16384,0.13500800397660997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,10240,0.02940622303220961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,10240,0.02405511173937056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,3072,0.028782222006056044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,3584,0.030108445220523413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,128,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,32,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,768,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,10240,0.03477599885728624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,2560,0.0070675553547011475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,6144,0.021433777279324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,1536,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,8192,0.013241777817408243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,7168,0.009660444325870937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,7168,0.04008444481425815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,2560,0.011332444846630096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,512,0.017243555850452848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,3072,0.015461333923869662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,768,0.0031999999450312722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,256,0.006224888894293044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,2560,0.012772444221708508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,1536,0.05302844444910685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,4096,0.3745680120256212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,5120,0.026018665896521673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,2560,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,8192,0.015124445160230001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,768,0.0035208890007601846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,1536,0.006732444382376141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,3072,0.014785778191354541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,128,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,2048,0.013878222141000958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,7168,0.1402497821384006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,65536,0.06593777736028035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,128,0.005225777626037598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,2560,0.004567999806669024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,1536,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,6144,0.009680888719028896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,128,0.012456888953844706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,256,0.002151111140847206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,768,0.009335999687512716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,4096,0.10446666346655951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,128,0.007070221834712558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,1536,0.018211555149820115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,1024,0.02604888876279195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,768,0.015126221709781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,768,0.00282577777074443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,12288,0.029876444074842665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,6144,0.009597332941161262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,768,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,6144,0.02643022272321913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,3072,0.2028426726659139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,2560,0.02500711050298479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,65536,0.17172000143263075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,3072,0.012960889273219638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,12288,0.028152889675564233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,4096,0.010336000058386061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,256,0.0024871110088295406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,256,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,3584,0.0052977779673205475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,1024,0.022318222456508215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,1536,0.013838222457302941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,3072,0.007237333390447829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,768,0.03456444541613261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,8192,0.010681777364677854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,16384,0.016551999582184684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,16384,0.021179555190934077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,6144,0.05888533592224121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,5120,0.44219289885626895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,512,0.025048888391918604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,5120,0.008344888687133789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,512,0.014085332552591959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,1536,0.01819644371668498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,512,0.002720000015364753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,6144,0.04953689045376248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,7168,0.012075555821259817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,768,0.006416888700591193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,6144,0.09627377986907959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,2560,0.004908444566859139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,1024,0.0028373334142896864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,32,0.0027919999427265594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,5120,0.007755555212497711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,256,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,65536,0.08380089203516643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,6144,0.013434666726324292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,1024,0.0038382220599386427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,32,0.004457777573002709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,7168,0.10268266995747884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,10240,0.016851555969980028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,5120,0.026380444566408794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,8192,0.025609778033362493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,5120,0.021583999196688335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,65536,0.030964444080988567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,10240,0.027618666489919026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,3584,0.007927999728255803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,8192,0.021850667066044275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,1536,0.01995644470055898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,65536,0.07141155666775174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,6144,0.029151111841201782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,10240,0.04440177811516655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,2560,0.02661688956949446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,256,0.015652444627549913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,10240,0.02094488839308421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,1024,0.00795022232664956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,5120,0.0914622214105394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,3072,0.007607111500369177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,512,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,12288,0.054923554261525474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,768,0.021007110675175984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,3584,0.0099982221921285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,768,0.014895111322402954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,64,0.001850666685236825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,10240,0.4701271057128906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,4096,0.009398221969604492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,512,0.014448000325096978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,128,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,3072,0.04589510957400004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,3072,0.007639110916190677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,768,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,6144,0.018526222970750596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,6144,2.338854259914822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,2048,0.019097778532240126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,768,0.00453955555955569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,12288,0.00796622203456031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,768,0.008593777815500895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,768,0.007928888830873702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,10240,0.03138577938079834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,128,0.013560888667901358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,12288,0.09464622206158108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,3584,0.01096088853147295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,2048,0.02960444490114848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,8192,0.010967111421955956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,7168,1.5417049196031358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,2560,0.01218311074707243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,6144,0.024920889072948035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,3072,0.01036088913679123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,4096,0.01849333279662662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,64,0.010975110861990186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,6144,0.03105955653720432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,1536,0.012096000214417776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,2560,0.004895110925038655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,3584,0.007682667010360294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,12288,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,1024,0.03082488973935445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,768,0.1221875614590115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,32,0.0021822222818930945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,256,0.009568889107969072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,6144,0.01846577723821004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,64,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,5120,0.02622311148378584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,5120,0.009292444421185387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,256,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,16384,0.022308443983395893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,12288,0.07464000251558092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,16384,0.07623644669850667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,65536,0.17977510558234322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,3584,0.005621333503060871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,16384,0.03250133328967624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,7168,0.03833422064781189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,7168,0.01794666714138455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,512,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,6144,0.03506311111980014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,3584,0.027640889088312786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,128,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,6144,0.0480275551478068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,3584,0.039332442813449435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,128,0.01014311114947001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,3072,0.012712889247470431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,12288,0.013424889081054263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,3584,0.004932444542646408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,1024,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,256,0.011347555451922946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,4096,0.011457777685589261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,512,0.002874666617976295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,6144,0.013083555632167392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,128,0.004524444540341695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,16384,0.10602133141623603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,2560,0.016166221764352586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,2048,0.006621333460013072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,128,0.0052968888647026485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,10240,0.03569066524505615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,3072,0.012703110774358114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,2560,0.08834844165378147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,4096,0.026280000805854797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,4096,0.015828443898095023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,512,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,7168,0.038252443075180054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,4096,0.010676444404655032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,128,0.02581866747803158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,256,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,16384,0.046160889996422656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,3072,0.006976000136799282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,16384,0.07134844197167291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,1536,0.039664000272750854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,3072,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,4096,0.0966035525004069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,1024,0.0037617778612507712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,1536,0.013103110922707452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,512,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,10240,0.04212444358401828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,2048,0.02089866664674547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,65536,0.03592533204290602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,2560,0.06588000059127808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,768,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,2560,0.01013777818944719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,4096,0.008383110993438298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,3072,0.03767911261982388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,6144,0.009638222555319468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,1024,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,16384,0.18954843944973418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,1536,0.006903999795516332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,32,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,65536,0.05724977784686618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,768,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,10240,0.025978666212823655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,256,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,3072,0.04368622104326884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,512,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,128,0.015154666370815702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,3584,0.015862221519152325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,2560,0.0069751110341813825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,128,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,10240,0.02736088964674208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,1536,0.020271999968422782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,128,0.014744000302420722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,256,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,3072,0.02366755571630266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,12288,0.013551111022631327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,5120,0.010415111151006486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,5120,0.03691822290420532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,3072,0.014130666851997375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,1024,0.04386311107211643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,32,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,7168,0.009974222216341231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,6144,0.018925334016482036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,32,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,12288,0.01959288948112064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,16384,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,16384,0.08874755435519749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,6144,0.018552000323931377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,12288,0.03107822272512648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,65536,0.03565688927968343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,12288,0.08159022198783027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,3072,0.014761777387724983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,16384,0.1503573391172621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,2048,0.00817599975400501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,6144,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,65536,0.07099911239412096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,5120,0.012472888661755456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,3584,0.013605333036846586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,1536,0.10429600212309097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,5120,0.0221742226017846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,768,0.015407999356587728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,10240,0.09825422366460164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,12288,0.2336275577545166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,12288,0.01847999956872728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,128,0.0018666666001081467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,7168,0.01036711119943195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,65536,0.05818666352166069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,768,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,6144,0.009319110876984065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,8192,0.011056000159846412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,7168,0.022319111559126113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,1024,0.010198222266303169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,4096,0.018231110440360177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,128,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,65536,0.09643377860387166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,12288,0.017879999346203275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,256,0.006565333240562015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,128,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,64,0.00180177783800496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,65536,0.12871021694607204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,512,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,768,0.011380444798204633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,16384,0.01613333324591319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,5120,0.008947555389669206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,3072,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,256,0.02040088839001126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,5120,0.01886311173439026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,256,0.0028640000770489373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,768,0.006544888847404056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,10240,0.02442488902144962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,1536,0.054013331731160484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,3584,0.0329075555006663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,1536,0.005547555370463266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,5120,0.040476444694730974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,6144,0.10886133379406399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,65536,0.04758400056097242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,128,0.002475555572244856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,2048,0.004368000146415499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,12288,0.013391999734772576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,5120,0.07669866747326322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,3584,0.2445324526892768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,16384,0.03613955444759793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,768,0.007970666719807519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,256,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,768,0.01757600075668759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,10240,0.12689956029256186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,1536,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,1536,0.006565333240562015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,3072,0.5509502092997233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,768,0.037680887513690524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,256,0.015445333388116626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,7168,0.012068444656001197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,6144,0.018219555417696636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,512,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,3584,0.04610933197869194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,12288,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,512,0.0028364445186323593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,10240,0.11527289284600152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,2048,0.01112711098459032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,12288,0.014703111516104804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,32,0.01955200069480472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,3584,0.007650666766696506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,3584,0.021923555268181696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,3584,0.007245333658324347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,2048,0.02056533263789283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,5120,0.008667555948098501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,65536,0.34268800417582196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,128,0.005251555393139522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,6144,0.00798044436507755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,16384,0.028910222980711196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,64,0.001793777777089013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,768,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,1536,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,8192,0.01959022217326694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,64,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,128,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,16384,0.015842666228612263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,16384,0.14195288552178278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,128,0.007649777664078607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,10240,0.016142222616407607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,64,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,128,0.013425778183672162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,1536,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,6144,0.036786665519078575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,3584,0.019371555911170114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,4096,0.03990755478541056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,128,0.014845333165592618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,7168,0.010758221977286868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,65536,0.19993333021799722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,2560,0.004940444396601783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,6144,0.03604266709751553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,65536,0.05952622493108114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,64,0.030841777722040813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,128,0.002436444370283021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,2048,0.011370666325092316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,6144,0.022984888818528917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,16384,0.02500088844034407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,1024,0.0031715554909573663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,12288,0.04574311110708448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,2560,0.00977955593003167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,256,0.013774221969975365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,256,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,12288,0.03152533372243246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,7168,0.01369599997997284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,1536,0.0690844456354777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,3584,0.025960000024901494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,65536,0.15018755859798855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,65536,0.059198220570882164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,8192,0.013653332988421122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,2048,0.02381155557102627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,1024,0.014119111829333834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,128,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,512,0.017282666431532968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,7168,0.02349066734313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,10240,0.04118577639261881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,5120,0.09933955801857842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,64,0.002152888932161861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,512,0.005584888988071018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,8192,0.023456000619464453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,16384,0.026940445105234783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,7168,0.021382222572962444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,65536,0.05049599872695076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,7168,0.018570666511853535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,65536,3.9874445597330728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,1024,0.006294222341643439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,6144,0.21704444620344374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,4096,0.020065777831607394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,12288,0.04392888810899523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,2048,0.006266666783226862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,2560,0.011043555206722684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,6144,0.010031999813185798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,1536,0.014465777410401238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,7168,0.06005511018964979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,512,0.002480000050531493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,6144,0.023551111419995625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,65536,0.02014222244421641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,256,0.015431111057599386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,1024,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,1024,0.012390221986505719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,8192,0.08650933371649848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,12288,0.030782222747802734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,768,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,2048,0.012402666939629449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,6144,0.05731911129421658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,128,0.014512888259357877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,3072,0.025015999873479206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,3072,0.013798221945762634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,64,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,4096,0.008984000318580205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,10240,0.02612888813018799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,16384,0.0410497784614563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,7168,0.010011555420027839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,512,0.01202666676706738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,5120,0.013049778011110095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,65536,3.601142035590278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,4096,0.018427555759747822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,4096,0.07804533508088854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,6144,0.008330666356616551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,12288,0.17807466453976104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,16384,0.036714666419559054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,128,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,5120,0.0122417774465349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,2560,0.012691555751694573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,3584,0.022632888621754114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,16384,0.33453067143758136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,2560,0.015832889411184523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,6144,0.008976000050703684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,2048,0.0045422220395671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,768,0.018852444158660043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,5120,0.009308444129096137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,128,0.01073155552148819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,2560,0.008311111066076491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,1024,0.02903288933965895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,16384,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,4096,0.008667555948098501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,6144,0.010033778018421596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,6144,0.009927999642160203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,7168,0.010878221856223213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,3584,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,1024,0.0310808883772956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,512,0.0031742221779293488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,5120,0.01924711134698656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,1024,0.01755288905567593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,3584,0.09877600272496541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,3584,0.009528888596428765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,128,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,10240,0.024819556209776137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,1536,0.015824889143308003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,5120,0.012144000165992312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,65536,0.12499111228519016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,10240,0.024422221713595923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,6144,0.33496710989210343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,7168,0.00958577791849772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,2048,0.014469333820872836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,3072,0.02129688858985901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,768,0.007250666618347168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,5120,0.02386222283045451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,768,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,7168,0.007648888561460707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,128,0.007589333587222629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,7168,0.00944800012641483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,512,0.0021546667234765156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,10240,0.03155733479393853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,64,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,3072,0.025920000341203477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,3072,0.008623110751310984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,2048,0.008763555851247575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,8192,0.03251377741495768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,512,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,8192,0.013459555804729462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,3584,0.025377778543366328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,512,0.0028195555011431375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,6144,0.008995555341243744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,5120,0.01346577786737018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,1024,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,7168,0.015096000499195524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,12288,0.034239110019471906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,64,0.0021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,3072,0.006909333169460297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,64,0.01092266705301073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,5120,0.02102222210831112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,2048,0.007065777977307637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,768,0.0030942221896515954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,12288,0.33054312070210773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,10240,0.01367644468943278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,2560,0.016167110866970487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,12288,0.060863110754224986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,3072,0.007276444799370236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,128,0.006646222124497096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,7168,0.00997511131895913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,65536,0.06633244620429145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,6144,0.03274577856063843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,3072,0.007616000043021307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,1536,0.008602666358153025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,65536,0.08810755279329087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,5120,0.016520889268981088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,2048,0.00793866647614373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,8192,0.013863110707865821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,12288,0.02188266648186578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,512,0.016456888781653512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,6144,0.012431110772821637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,10240,0.016182222300105624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,3584,0.015591111448076038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,8192,0.043888890080981784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,3584,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,7168,0.03763733307520548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,2560,0.011319110790888468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,32,0.0052693333062860705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,64,0.009295110901196798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,256,0.017318222257826064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,3072,0.01585244470172458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,256,0.0021173333128293357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,16384,0.042436444097095065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,768,0.00630044440428416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,7168,0.017733333839310538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,256,0.004870222260554631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,4096,0.0966871115896437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,2560,0.007244444555706448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,8192,0.025062221619817946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,7168,0.02145066691769494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,2560,0.014815110299322339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,65536,0.08126399914423625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,512,0.0025475554996066624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,12288,0.03472355670399136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,4096,0.03714933329158359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,16384,0.08163288566801283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,4096,0.024041778511471216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,768,0.0028080000645584534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,3584,0.02233244478702545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,512,0.025292444560262892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,4096,0.015862221519152325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,512,0.01586577792962392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,16384,2.3594932556152344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,8192,0.022268444299697876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,256,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,256,0.005438222239414851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,512,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,65536,0.12303911315070258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,8192,0.022135999467637803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,2560,0.014019555515713163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,256,0.005947555518812603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,6144,0.011711999773979187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,512,0.019319110446506076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,1536,0.006227555374304454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,8192,0.01796800063716041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,65536,1.4858764012654622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,3072,0.005909333212508096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,128,0.002510222295920054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,5120,0.009308444129096137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,256,0.0064560001095136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,8192,0.013243555194801755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,8192,0.04923466510242886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,16384,0.07953866985109118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,16384,0.03244888782501221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,1024,0.014431110686726041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,64,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,32,0.001814222170246972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,3072,0.00759733302725686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,2048,0.11428622404734294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,65536,0.009144888983832465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,128,0.0058755555914507965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,16384,0.028367112080256145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,128,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,4096,0.015248000621795654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,2560,0.009573332965373993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,64,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,65536,0.09971111350589329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,64,0.0035982223020659555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,512,0.0028204443968004654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,12288,0.060170663727654346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,6144,0.027004443936877783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,5120,0.045816888411839805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,768,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,10240,0.02060622142420875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,7168,0.1716302235921224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,12288,0.03975111246109009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,3584,0.013857777747843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,2560,0.006803555621041193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,1536,0.0038817777401871155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,32,0.0028684443483750024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,65536,0.44127289454142254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,12288,0.01717333330048455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,6144,0.04731733269161648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,6144,0.010888000329335531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,8192,0.14742755889892578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,16384,0.022980444961124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,32,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,6144,0.012124444047609964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,256,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,10240,0.01088177743885252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,6144,0.005203555441564984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,64,0.005974222388532426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,16384,0.04931022061241997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,128,0.0021582223061058256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,512,0.003181333343187968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,10240,0.4556115468343099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,768,0.012489777472284106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,2560,0.014675555957688225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,3584,0.025311999850802954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,3584,0.012639111114872826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,2560,0.012785777449607849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,5120,0.0083137775460879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,768,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,2560,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,1536,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,32,0.0021893332401911416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,32,0.0028435554769304064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,3584,0.007721777591440413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,6144,0.011137777732478248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,3072,0.008656889200210571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,5120,0.025640888346566096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,12288,0.020495999190542433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,3584,0.011018666956159802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,8192,0.4638204574584961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,1024,0.0032257777121331957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,256,0.003630222131808599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,128,0.00796888851457172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,32,0.00319377767542998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,10240,0.0326533317565918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,512,0.07000178098678589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,4096,0.01579466627703773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,3584,0.0949911144044664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,6144,0.034191999170515276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,768,0.005635555419656966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,128,0.006986666884687211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,2560,0.012439111040698158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,2048,0.02810933192571004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,32,0.002054222135080232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,7168,0.010512000156773461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,7168,0.016148444679048326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,2048,0.00676711110605134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,12288,0.02881866693496704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,3072,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,6144,0.016876444220542908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,10240,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,16384,0.024728000164031982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,4096,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,65536,0.12817600038316515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,2560,0.02370311154259576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,1024,0.0038497779104444715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,512,0.005149333427349727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,6144,0.011873777541849347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,32,0.013122667041089801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,10240,0.02572000026702881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,3584,0.008934222161769867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,16384,0.03529689047071669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,2048,0.012444444828563266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,8192,0.50910398695204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,1536,0.02162844439347585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,1024,0.005958222266700532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,64,0.006806222101052602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,2560,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,128,0.015116444892353482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,6144,0.013174222575293647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,65536,0.24729156494140625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,10240,0.08619911140865749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,512,0.0025235555238193935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,6144,0.017439110411538016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,10240,0.015495111544926962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,3072,0.026955554882685345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,8192,0.018183110488785636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,6144,0.2943626774681939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,7168,0.009865778187910715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,256,0.027874665127860174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,1024,0.022107554806603327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,12288,1.0406746334499783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,3584,0.007752888732486301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,2560,0.00488800017370118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,12288,0.03235022226969401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,2048,0.018532445033391316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,3584,0.028096887800428603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,1024,0.017807111144065857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,32,0.0028311111446883944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,2560,0.009350222018029954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,512,0.0064106666379504735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,16384,0.027573333846198186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,1536,0.012724444270133972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,3584,0.004861333303981357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,16384,0.023966223001480103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,65536,0.05901777744293213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,3072,0.1345155504014757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,16384,0.020896888441509668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,10240,0.09745599826176961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,6144,0.1140257782406277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,3584,0.013624000052611033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,5120,3.769891526963976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,5120,0.01384888920519087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,8192,0.015811555915408664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,1024,0.015280000037617154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,16384,0.04772444566090902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,10240,0.04308355516857571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,10240,0.014102222190962898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,256,0.002465777720014254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,3584,0.015506666567590503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,1536,0.013783111340469785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,3072,0.024131556351979572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,4096,0.029913776450686987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,512,0.005927111125654644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,3072,0.031472888257768415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,768,0.5779813130696615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,7168,0.009102221992280748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,2560,0.016160888804329764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,4096,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,512,0.00664533343580034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,256,0.018586667047606576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,10240,0.18685332934061685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,3584,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,3072,0.00906311141120063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,3072,0.018726223044925265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,2048,0.24679999881320527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,3584,0.01641155613793267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,512,0.002883555574549569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,64,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,1024,0.012019555601808759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,10240,0.008656889200210571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,3584,0.008200888832410177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,7168,0.01886399918132358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,5120,0.008641777767075432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,768,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,2048,0.00666133314371109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,4096,0.02176711128817664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,2048,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,10240,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,65536,0.061667554908328585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,4096,0.012256888879670037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,65536,0.11769955688052708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,128,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,5120,0.04689866635534498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,7168,0.013437333206335703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,128,0.002163555473089218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,2048,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,5120,0.01349955548842748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,65536,0.02368444369898902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,512,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,5120,0.006561777657932705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,3072,0.01072177787621816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,128,0.0018542222678661346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,16384,0.09645422299702962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,512,0.012792888614866467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,6144,0.1697822279400296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,1024,0.015840000576443143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,768,0.006552888701359431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,2048,0.006728000111050076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,3072,0.011401777466138205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,7168,0.012517333030700684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,2048,0.015106666419241162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,5120,0.033703111939960055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,1024,0.0031431110368834603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,65536,0.08800533082750107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,2048,0.006679111056857639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,256,0.005212444398138259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,8192,0.020945777495702107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,7168,0.028451555305057104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,1024,0.013415111435784234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,8192,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,16384,0.040511998865339495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,1536,0.010207999911573198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,8192,0.005497777627574072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,3584,0.06684177451663546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,512,0.013106666505336761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,4096,0.009687110781669617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,1024,0.014147554834683737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,64,0.001847111102607515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,1536,0.019895111521085102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,1536,0.0038453332251972626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,4096,0.00453333349691497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,2048,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,8192,0.07662844657897949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,6144,0.23607911003960502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,768,0.26285600662231445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,512,0.012438221938080259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,3584,0.012849777936935425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,1536,0.016499555773205228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,64,0.013072888884279462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,2048,0.004545777622196409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,16384,0.06569511360592313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,65536,0.0628897746404012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,128,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,65536,0.04881777697139316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,6144,0.019172444939613342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,12288,0.03276711040072971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,128,0.002510222295920054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,16384,0.020706666840447318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,768,0.03772799836264716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,12288,0.34357688162061906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,1536,0.006139555739031897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,4096,0.02217688825395372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,10240,0.022248889009157818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,7168,0.010236444572607676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,1024,0.008642666869693333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,5120,0.03320533368322585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,768,0.0031715554909573663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,256,0.030208888981077407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,3584,0.005626666463083691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,64,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,6144,0.011113777756690979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,16384,0.026514665948020086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,5120,0.021957332889238994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,3584,0.20158667034573027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,10240,0.026449778013759192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,7168,0.13965688811408147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,2560,0.009667555491129557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,6144,0.016504888733228047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,512,0.02197066611713833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,128,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,256,0.015135111080275642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,3072,0.06017422013812595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,5120,0.02014666630162133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,4096,0.02496533261405097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,1024,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,4096,0.031168888012568157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,4096,0.05322133170233833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,8192,0.008120888637171851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,5120,0.008440000315507254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,5120,0.03038311004638672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,6144,0.017131555411550734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,128,0.009024000002278222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,3072,0.009663110805882348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,3072,0.007900444169839224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,3584,0.013651555611027611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,6144,0.02156266735659705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,1024,0.0038480001191298165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,768,0.02197155521975623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,256,0.002151111140847206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,65536,0.320266670650906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,1536,0.019966221517986722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,7168,0.020280889338917203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,256,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,5120,0.025254221426116094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,768,0.0063235556913746735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,2560,0.007288888924651676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,16384,0.020602666669421725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,256,0.0022168887986077205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,7168,0.010695111420419482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,2048,0.012401777837011548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,7168,0.009312888814343346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,1536,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,768,0.006551110910044775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,65536,0.06085244152281019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,3072,0.007275555696752336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,768,0.01756622228357527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,8192,0.011006222003036074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,6144,0.04011822077963088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,512,0.00655733338660664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,1536,0.003984000119898054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,64,0.007600000335110559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,10240,0.01629599928855896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,128,0.0052693333062860705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,64,0.010776888993051318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,8192,0.03222577770551046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,65536,0.024938666158252295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,10240,0.026879111925760906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,10240,0.012521777715947894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,2560,0.02092444399992625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,256,0.013045333325862885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,5120,0.2780666616227892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,10240,0.28670043415493435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,12288,0.014133334159851074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,2560,0.013117333253224691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,128,0.01514933341079288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,512,0.014130666851997375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,5120,0.006225777582989798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,512,0.012489777472284106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,7168,0.009649777577983009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,2560,0.01789422167672051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,6144,0.10162488619486491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,768,0.01852888862291972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,64,0.002855111120475663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,3584,0.006225777582989798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,32,0.005533333453867171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,6144,0.007576888634098901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,12288,0.06559644142786662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,768,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,4096,0.08601599931716919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,4096,0.014278221461508008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,2048,0.009319999979601966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,2048,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,8192,0.020625778370433383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,3584,0.0406915545463562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,3584,0.012896888785892062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,256,0.014799111419253878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,3584,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,8192,0.3618684344821506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,768,0.0029893333299292457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,32,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,1024,0.007631111476156447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,128,0.013787555197874704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,2560,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,512,0.006638222270541721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,3072,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,32,0.006224888894293044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,10240,0.012471111284361945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,7168,0.07851111226611666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,256,0.0021297778520319196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,32,0.0069288888739215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,2048,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,768,0.0028684443483750024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,4096,0.09842222266727024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,16384,0.03382844395107693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,3072,0.006266666783226862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,768,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,65536,0.08227111233605279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,6144,0.03028622269630432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,65536,0.17518399821387398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,5120,0.020277332928445604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,4096,0.008435555630260045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,6144,0.013617777989970313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,3072,0.00742222203148736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,256,0.015849777393870883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,256,0.004902222090297275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,512,0.014126222994592456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,8192,0.040376888381110296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,256,0.016495111915800307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,5120,0.009977777798970541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,64,0.0038613333470291565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,16384,0.013551111022631327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,6144,0.009312888814343346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,10240,0.006903111106819577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,10240,0.020847999387317236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,3072,0.007635555333561367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,3584,0.03473688827620612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,4096,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,1536,0.02277511192692651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,10240,0.024920889072948035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,6144,0.24225687980651855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,4096,0.014470222923490735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,16384,0.016140444411171805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,256,0.018551111221313477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,6144,0.026868444350030687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,2560,0.06822222471237183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,3584,0.008294222255547842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,10240,0.027076444692081873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,512,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,3072,0.01037333326207267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,65536,0.16971645090315077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,768,0.01553600033124288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,2560,0.015811555915408664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,128,0.00554044461912579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,2048,0.08276000287797716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,1024,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,12288,0.030424889591005113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,10240,0.013450666434235044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,10240,0.013403555585278405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,1024,0.013449778159459433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,1024,0.03390489021937052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,512,0.0032151111712058387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,7168,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,768,0.005715555614895291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,4096,0.06403822369045682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,2560,0.008150222400824228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,768,0.056448002656300865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,32,0.005249777601824866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,2048,0.02123822271823883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,2048,0.01648177703221639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,7168,0.009343999955389235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,32,0.0053031109273433685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,1024,0.00629333323902554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,65536,0.07915911409589979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,16384,0.03914311197068956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,2048,0.0069573331210348345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,6144,0.039464887645509504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,16384,0.016520889268981088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,10240,0.014902222487661572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,3072,0.007256888680987888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,5120,0.010371555884679159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,512,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,1536,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,3072,0.007619555625650618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,512,0.003167111012670729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,3584,0.009200000100665623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,3584,0.05989244249131945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,10240,0.022482666704389784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,10240,0.19456444846259224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,1536,0.011724444727102915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,8192,0.015172445111804538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,10240,0.009089777866999308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,7168,0.19844978385501436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,3072,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,3072,0.01683911184469859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,65536,0.04746311240726047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,16384,0.03659111261367798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,12288,1.3008960088094075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,2560,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,10240,0.03715733355946011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,5120,0.02809600035349528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,64,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,16384,0.020988444487253826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,512,0.007754666937722101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,2048,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,128,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,7168,0.13979732990264893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,2048,0.021928000781271193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,5120,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,1024,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,768,0.012057777908113269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,7168,0.379314661026001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,256,0.0021884443445338146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,5120,0.008989333278603023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,512,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,512,0.003192000091075897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,8192,0.10860088798734878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,1536,0.0032764443506797156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,10240,0.019511999355422128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,3072,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,7168,0.3710622257656521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,256,0.004844444493452708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,3072,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,1024,0.04868355393409729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,256,0.007019555403126611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,10240,0.025120000044504803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,128,0.09016799926757812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,4096,0.008999110923873054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,16384,0.03144977821244134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,3584,0.004528000122971005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,768,0.005883555445406172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,3072,0.01971733404530419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,6144,0.012757333616415659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,512,0.015545777148670621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,512,0.002483555633160803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,64,0.007988444632954067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,64,0.002489777695801523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,64,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,8192,0.06004089117050171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,2560,0.0058319999112023254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,128,0.005177777674463061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,12288,0.026707554856936138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,6144,0.013582222163677216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,128,0.005377777748637729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,12288,0.1321671141518487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,16384,0.014476444986131458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,64,0.04058666692839728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,16384,0.03017688790957133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,256,0.0024497778051429326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,32,0.005631111148330901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,5120,0.05848177936342028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,6144,0.018592889110247295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,512,0.0174319992462794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,4096,0.007988444632954067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,16384,0.022112000319692824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,5120,0.013510222236315409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,768,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,512,0.0025128889828920364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,5120,0.03082488973935445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,1024,0.005967999911970562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,64,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,3072,0.004583111239804162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,2560,0.0069217777086628815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,128,0.02128977742460039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,65536,3.6932576497395835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,768,0.010375111467308467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,8192,0.007974222302436829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,6144,0.015957333975368075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,768,0.027479110492600337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,512,0.006963555183675554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,64,0.001840888833006223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,4096,0.008273777862389883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,2048,0.007853333320882585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,64,0.00249422217408816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,65536,0.24284089936150444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,6144,0.03827111257447137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,3072,0.007657777931955125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,3584,0.014440000057220459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,16384,0.08046311140060425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,12288,1.0478444629245334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,1024,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,512,0.016505777835845947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,1536,0.005527110977305307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,8192,0.11382222175598145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,3072,0.010686222049925061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,16384,0.0325271122985416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,1536,0.015122666954994202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,4096,0.011376889215575324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,65536,0.07656000057856242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,7168,0.03162666824128892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,10240,0.03743288914362589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,1536,0.005964444329341252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,1536,0.0039057777159743835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,3584,0.07872977521684435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,1536,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,7168,0.03568089008331299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,2560,0.016215999921162922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,5120,0.031088001198238794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,3072,0.6572364701165093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,7168,0.010049777726332346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,3072,0.018953777021831937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,64,0.004896000027656555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,3072,0.008047999607192146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,4096,0.03365955419010586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,128,0.014486221803559197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,16384,0.02422399984465705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,5120,0.012167111039161682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,128,0.022833777798546687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,65536,0.04771911104520162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,16384,0.033371554480658636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,16384,0.024859555893474158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,1024,0.02516355613867442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,768,0.02257777750492096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,8192,0.009157333109113906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,1536,0.005982222242487802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,128,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,512,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,65536,0.07792622513241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,1536,0.00518133325709237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,10240,0.02498222225242191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,3072,0.007254222200976477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,12288,0.00508088908261723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,1536,0.0388995541466607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,65536,0.06416800287034777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,12288,0.00490044429898262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,16384,0.05102133419778612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,12288,0.02667733364635044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,1536,0.0459591117170122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,4096,0.015240889456537036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,4096,0.021959111094474792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,32,0.001826666709449556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,10240,0.00867022242810991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,10240,0.10815911160575019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,16384,0.0812986691792806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,3072,0.02287111183007558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,1024,0.008358221915033128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,256,0.006653333289755716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,2048,0.01992266707950168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,2560,0.005728000154097875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,2560,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,3072,0.00740444411834081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,2048,0.0075733330514695905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,65536,0.7334551281399198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,12288,0.22117866410149467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,512,0.009004444711738164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,1024,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,256,0.005236444373925527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,768,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,3072,0.02660266723897722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,128,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,12288,0.018248000078731112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,128,0.00629333323902554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,256,0.009316444396972656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,4096,0.007996444072988298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,128,0.017701332767804463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,3072,0.013079110946920184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,7168,0.01109866632355584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,8192,0.1191448900434706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,32,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,768,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,3584,0.021293333835071985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,1024,0.014977777997652689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,10240,0.014427555931939019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,3072,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,128,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,256,0.0025111111915773815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,10240,0.012151111331250934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,1536,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,3072,0.007264888948864407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,768,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,256,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,10240,0.03426488902833726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,16384,0.027504000398847792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,7168,0.019869332512219746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,512,0.009013333254390294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,2560,0.027328888575236004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,65536,0.27530399958292645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,512,0.013172444370057849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,3072,0.0053431110249625305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,1024,0.005776888794369168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,6144,0.010601777997281818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,16384,0.020948444803555805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,4096,0.00794488853878445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,6144,0.035869333479139544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,32,0.0021199999998013177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,3072,0.019627556204795837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,16384,0.046450667911105685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,3072,0.01477688882086012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,8192,0.01460977726512485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,7168,0.021241777473025854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,256,0.005222222043408289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,1024,0.007605333295133378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,65536,0.053505778312683105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,3072,0.014920888675583733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,1024,0.008603555460770925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,1536,0.031911111540264554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,6144,0.014454222387737699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,16384,0.020248000820477802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,4096,0.0685066646999783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,64,0.004864000197913912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,128,0.014814222852389017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,12288,0.04917866653866238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,8192,0.014640889234013028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,768,0.0041742221348815495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,5120,0.027678221464157104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,6144,0.012539555629094442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,64,0.0058764442801475525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,10240,0.059437334537506104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,128,0.012832000023788877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,65536,0.06638755400975545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,256,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,128,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,7168,0.020032000210550096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,65536,0.2266062100728353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,3072,0.007596444752481248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,6144,0.020384889509942796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,2560,0.019897777173254225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,7168,0.018232888645595975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,6144,0.015326221783955893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,2560,0.007307555940416124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,4096,0.016855110724767048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,7168,0.009703111317422655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,2048,0.015136889285511441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,10240,0.34133511119418675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,65536,3.037531534830729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,10240,0.38254398769802517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,32,0.004228444563017951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,768,0.002521777732504739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,7168,0.008040889269775814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,2048,0.007308444215191736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,10240,0.1358542177412245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,64,0.008301333420806462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,768,0.031699554787741765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,8192,0.035818666219711304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,8192,0.025506666964954797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,65536,0.20631110668182373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,32,0.002452444492114915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,32,0.026595556073718604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,32,0.01276088919904497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,768,0.004538666870858935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,65536,0.38041244612799746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,3584,0.1535093386967977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,2048,0.011325333681371478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,128,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,7168,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,32,0.012811555630630918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,65536,0.011380444798204633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,3072,0.0630355543560452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,8192,0.030241777499516804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,16384,0.0199644449684355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,8192,0.06977866755591498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,5120,0.010996444357766045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,6144,0.015805333852767944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,1536,0.01424088908566369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,2560,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,128,0.019543111324310303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,768,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,512,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,256,0.0034640000926123727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,10240,0.03305600086847941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,2048,0.0069395556218094295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,1536,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,7168,0.031001776456832886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,512,0.010039110978444418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,16384,0.17326933807796904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,2560,0.007240888973077138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,64,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,768,0.002471111093958219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,10240,0.015055111712879606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,65536,0.4653244548373752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,2560,0.004882666799757215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,1024,0.006486222147941589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,10240,0.015156444576051501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,128,0.01000355515215132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,64,0.007649777664078607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,7168,0.027066666218969557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,12288,0.016564443707466125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,8192,0.03225244416130914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,65536,0.11038133170869614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,1536,0.006245333287451003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,256,0.002141333288616604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,12288,0.02568888829814063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,7168,0.17814221647050646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,4096,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,3584,0.01479288935661316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,5120,0.01202044470442666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,32,0.005531555662552516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,128,0.0023431111541059283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,16384,0.0318248901102278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,65536,0.00656088896923595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,1024,0.0035297779573334586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,10240,0.016620443926917184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,32,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,3584,0.009877333210574256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,64,0.0073902226156658595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,8192,0.010860444770918952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,6144,0.012300444145997366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,8192,0.027503111296229895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,65536,0.07433333661821154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,12288,0.01940355532699161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,64,0.0052471111218134565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,768,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,2048,0.009722666607962714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,256,0.015110222829712762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,3072,0.01774133410718706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,1536,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,256,0.006590222318967183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,10240,0.37553956773546004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,3072,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,10240,0.0310479998588562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,6144,0.011743110915025076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,2560,0.01111644423670239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,3072,0.02125688890616099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,1536,0.009322666459613377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,2560,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,16384,0.07107466459274292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,1024,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,4096,0.021996445126003687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,65536,0.08461333645714654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,2560,0.006676444576846228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,5120,0.025729777084456548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,10240,0.02433066732353634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,3072,0.007263110743628607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,256,0.023021333747439917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,8192,0.0192284451590644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,8192,0.03035022152794732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,7168,0.027760000692473516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,2048,0.004569777597983678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,128,0.005304000029961268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,4096,0.018183999591403537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,8192,0.016512889001104567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,65536,0.07567644119262695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,16384,0.006531555619504716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,5120,0.05996533234914144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,16384,0.01734133395883772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,768,0.003499555504984326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,8192,0.14712266127268472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,4096,0.011861333416567909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,3584,0.014951111541854011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,4096,0.03458844290839301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,256,0.018206222189797294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,768,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,256,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,3584,0.02985066506597731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,32,0.002187555655837059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,4096,0.008280889027648503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,2048,0.009930666950013902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,128,0.01480888823668162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,10240,0.01724177764521705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,2048,0.013807111316257052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,768,0.0028115556471877625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,65536,0.19422222508324516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,10240,0.012337778177526263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,5120,0.023507555325826008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,12288,0.024187554915746052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,12288,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,4096,0.02478133307562934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,3072,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,2048,0.02880355715751648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,8192,0.03225333491961161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,8192,0.13707821898990208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,65536,0.12179288599226211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,12288,0.06363377968470256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,7168,0.007855999800893996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,3584,0.006092444476154115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,768,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,2560,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,10240,0.0229777776532703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,768,0.0038746665749284956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,8192,0.021988444858127173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,1536,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,6144,0.04725066820780436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,128,0.0021155555215146807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,2048,0.00798044436507755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,10240,0.021292444732454088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,12288,0.048017776674694486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,1024,0.013774221969975365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,6144,0.015860444969601102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,4096,0.01108888867828581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,6144,0.03902755512131585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,1536,0.01274222218328052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,3072,0.022630222969584998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,3584,0.037667555941475764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,128,0.002893333426780171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,16384,0.27476443184746635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,2560,0.007247111035717859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,65536,0.07444710863961114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,1536,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,12288,0.03378933336999681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,4096,0.03638489047686259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,1536,0.007432888779375289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,65536,0.12391822867923313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,6144,0.014088888963063559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,3072,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,3584,0.006657777975002925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,65536,0.13522666030459934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,1536,0.12371199660831028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,16384,0.02977422210905287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,8192,0.005665777872006099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,10240,0.03237422307332357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,16384,0.022590221630202398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,2048,0.0123511114054256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,3584,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,10240,0.06590044498443604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,6144,0.023235556152131822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,1024,0.012744888663291931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,10240,0.012092444631788464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,1024,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,2560,0.023590222001075745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,16384,0.037900444534089826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,2048,0.03907555672857497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,4096,0.1090080009566413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,1536,0.00665244460105896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,256,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,768,0.01590399940808614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,768,0.006897777732875612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,6144,0.007547555698288812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,12288,0.02913244565327962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,5120,0.1756488879521688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,256,0.01514755520555708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,8192,0.025018667181332905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,64,0.011695110963450538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,128,0.0103493332862854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,10240,0.03691822290420532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,512,0.002499555548032125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,12288,0.1049928863843282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,2048,0.008967111508051554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,12288,0.017490666773584154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,8192,0.010670222342014313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,65536,0.18260978327857125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,512,0.015115555789735583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,8192,0.07454133033752441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,64,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,12288,0.03787911269399855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,2048,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,256,0.010012444522645738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,3072,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,128,0.0021191111041439907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,768,0.005863111052248213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,6144,0.01219555570019616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,3584,0.0669635534286499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,1536,0.01718311177359687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,16384,0.02390044430891673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,64,0.10214755270216201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,256,0.01479822231663598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,65536,0.10783644517262776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,256,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,6144,0.04972889026006063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,2048,0.021547555923461914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,64,0.0018639999131361644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,256,0.015483554866578845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,128,0.004916444420814514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,256,0.019909333851602342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,3072,0.014304889572991265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,256,0.01646933290693495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,5120,0.07936444547441271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,256,0.02495555579662323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,768,0.005632888939645555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,4096,0.009128889275921715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,2560,0.013088888592190213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,512,0.009492444495360056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,256,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,1536,0.02179555594921112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,512,0.06504888667000665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,3072,0.015652444627549913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,65536,0.8002550866868762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,512,0.01034577770365609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,3072,0.07189066542519464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,16384,0.0381991101635827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,5120,0.028839998775058325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,65536,0.1049742235077752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,6144,0.011042666931947073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,16384,0.0222746663623386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,3072,0.055163555675082736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,65536,0.33026221063401967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,2560,0.007118222614129384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,1024,0.008344888687133789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,1536,1.0628195868598092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,6144,0.021872889664438035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,32,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,8192,0.014644443988800049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,4096,0.013843555417325763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,512,0.004893333133724001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,256,0.0063440000845326324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,512,0.007320889168315464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,2560,0.00923555592695872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,4096,0.007807110746701558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,1024,0.00962400022480223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,3072,0.08573333422342937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,768,0.0031591111587153543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,32,0.001793777777089013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,3072,0.007288888924651676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,12288,0.05282310975922478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,2560,0.011514667007658215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,5120,0.012840888566441007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,2560,0.02653955585426754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,128,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,768,0.017154667112562392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,3584,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,4096,0.030745777818891738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,2560,0.020599111914634705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,128,0.01923022170861562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,8192,0.046401778856913246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,10240,0.018004443910386827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,1024,0.007510222494602203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,2560,0.013020444247457715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,8192,0.023669333921538457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,128,0.05935021903779772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,6144,0.017165333032608032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,8192,0.01071111112833023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,512,0.016977777083714802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,65536,0.10614044136471218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,10240,0.18109688493940565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,2048,0.004881777697139316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,5120,0.08615289131800334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,768,0.018174222773975797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,128,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,3072,0.009381333159075843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,512,0.0028168888141711554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,3584,0.014135999812020196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,12288,0.013449778159459433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,12288,0.1345306634902954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,768,0.0140666663646698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,3584,0.022095110681321885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,2048,0.011386666860845355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,12288,0.04951111144489712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,2048,0.0068986668354935115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,256,0.015460444821251763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,5120,0.017495999733606975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,3072,0.03566933340496487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,64,0.0018240000224775737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,5120,0.033921778202056885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,1536,0.020720889170964558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,512,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,128,0.019559999306996662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,10240,0.06581066714392768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,3584,0.00829777783817715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,12288,0.03484355409940084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,512,0.004190222256713443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,7168,0.04554933309555054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,3072,0.007245333658324347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,512,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,7168,0.01793511046303643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,10240,0.0184897780418396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,2560,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,1536,0.01088355564408832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,12288,0.041054222318861223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,2048,0.004499555461936527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,256,0.01682222220632765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,256,0.00832622249921163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,4096,0.009843555589516958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,128,0.006207110981146495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,65536,0.0760293338033888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,64,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,1024,0.006303999986913469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,512,0.03048355711830987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,1536,0.0058782220714622075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,1024,0.00795555528667238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,2560,0.007032000356250339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,16384,0.03207644489076402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,4096,0.00834222220712238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,1024,0.011737777955002256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,8192,0.010464888479974536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,16384,0.34594400723775226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,128,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,16384,0.09614755709966023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,65536,0.07757777637905545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,512,0.003863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,4096,0.00977955593003167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,65536,0.33276711569892037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,16384,0.02281422250800663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,768,0.0028213332924577924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,10240,0.030842665168974135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,64,0.0018195555441909367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,1536,0.006227555374304454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,5120,0.04625422093603346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,6144,0.00965244405799442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,256,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,16384,0.06215644545025296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,12288,0.014075555735164218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,12288,0.016873778568373788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,5120,0.011624000138706632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,128,0.016140444411171805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,5120,0.018180444836616516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,65536,0.0913626684082879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,5120,0.03747111227777269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,768,0.009328888522254096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,768,0.04079111085997687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,2560,0.006685333533419504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,128,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,12288,0.03596622082922194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,6144,0.005224888937340842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,5120,0.02367288867632548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,16384,0.04301066531075371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,2048,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,512,0.013435555828942193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,1024,0.014834667245546976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,10240,0.013088888592190213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,3072,0.15402666727701822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,5120,0.009332444104883406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,4096,0.011739555332395764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,10240,0.4193280008104112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,5120,0.03514488869243198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,6144,0.2627057764265272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,3584,0.028049776951471966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,65536,0.06653689013587104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,7168,0.025507556067572698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,3072,0.01756000022093455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,1536,0.043360888957977295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,10240,0.014135110709402295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,3584,0.009156444834338294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,3584,0.011365333365069495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,7168,0.029808001385794744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,1024,0.014124444789356656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,7168,0.00795999997191959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,512,0.014133334159851074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,8192,0.029882666137483384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,128,0.0021582223061058256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,5120,0.01889600025282966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,4096,0.017530666457282174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,8192,0.022125333547592163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,6144,0.027168888184759352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,3584,0.016502221425374348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,3072,0.08492355876498753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,5120,0.00904266701804267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,1024,0.007268444531493717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,2560,0.022692445251676772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,6144,0.05454311105940077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,10240,0.055528889099756874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,3072,0.012072888513406118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,8192,0.05018844538264804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,2048,0.007296889192528195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,1024,0.03085422184732225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,768,0.016806221670574613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,10240,0.025478222303920325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,8192,0.010703110860453712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,2048,0.019896000623703003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,16384,0.008307555483447181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,32,0.001814222170246972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,2048,0.009744889206356471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,32,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,768,0.018181333939234417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,16384,0.0579964452319675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,8192,0.012434666355450949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,4096,0.013791110780504016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,768,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,128,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,512,0.022840888963805303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,2560,0.014905777242448596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,7168,0.010989333192507425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,65536,0.16540711455874973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,16384,0.005616888817813661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,3072,0.01651733285850949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,2560,0.007314666277832455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,768,0.019633778267436557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,512,0.0058782220714622075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,6144,0.01050222251150343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,8192,0.011015111373530494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,12288,0.048355556196636625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,3072,0.02973066767056783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,16384,0.01959377858373854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,1536,0.0063822223908371395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,65536,0.16665244102478027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,2048,0.02782400117980109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,16384,0.022750222020679053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,65536,0.34633776876661515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,12288,0.07024977604548137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,1024,0.006229333165619109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,10240,0.02132622235351139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,6144,0.014392000105645923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,1024,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,16384,0.041675554381476514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,8192,0.07467289103402032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,16384,0.017676444517241586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,512,0.023010666171709698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,768,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,8192,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,1536,0.008291555775536431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,1024,0.0075866662793689305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,6144,0.04486044579082065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,1024,0.012104888757069906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,12288,0.029626667499542236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,128,0.0140684445699056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,10240,0.04759822289148966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,3584,0.007615110940403408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,7168,0.2785573270585802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,16384,0.034048000971476235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,768,0.0028195555011431375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,2048,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,8192,0.012102222277058495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,6144,0.00998755544424057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,1536,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,1024,0.013046222428480784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,2560,0.0748035576608446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,1024,0.021469333105617102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,16384,0.016480889585283067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,3584,0.009753777749008602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,3072,0.03642488850487603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,5120,0.02074666652414534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,1024,0.004556444370084339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,256,0.014496000276671516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,128,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,128,0.002102222293615341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,65536,0.058541331026289195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,32,0.007118222614129384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,3072,0.012665777570671506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,10240,0.006680000159475539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,3584,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,1024,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,2560,0.0052266667286554975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,32,0.0021119999388853707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,1024,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,1536,0.003488000068399641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,2048,0.01034400032626258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,5120,0.016157334049542744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,10240,0.017111111018392775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,1536,0.015513777732849121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,16384,0.03356711069742838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,65536,0.05673778057098389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,10240,0.05605866511662801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,1024,0.006441777778996362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,16384,0.03631022241380479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,512,0.005184000151024925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,1536,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,3584,0.010272000398900773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,6144,2.1985840267605252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,16384,0.04274310999446445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,64,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,2560,0.008564444051848518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,1024,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,512,0.014196443888876172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,1536,0.003555555724435382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,768,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,3584,0.007601777712504069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,6144,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,12288,0.041089776489469744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,512,0.0028293333533737394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,2048,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,7168,0.6201875474717882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,768,0.023520888553725347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,65536,0.013342222405804528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,3072,0.009959999885823991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,256,0.003855111284388436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,768,0.0554986662334866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,3072,0.022551111049122278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,1024,0.033723556333118014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,3072,0.014154665999942355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,32,0.0018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,7168,0.01629955569903056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,512,0.015138667490747241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,64,0.003850666599141227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,10240,0.5264151361253526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,2048,0.10008177492353652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,1536,0.006933333145247565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,5120,0.026855111122131348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,256,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,32,0.004885333279768626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,1536,0.0032186667538351486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,4096,0.10673155387242635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,8192,0.020270221763186984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,2048,0.005166222237878376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,8192,0.004219555606444676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,10240,0.04252800014283922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,2560,0.11833511458502875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,3072,0.012358222570684222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,64,0.0021724444296624926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,2048,0.08951110972298516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,6144,0.011055111057228513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,3072,0.016508445143699646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,8192,0.01889155473974016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,6144,0.10538933012220596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,10240,0.008926221893893348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,6144,0.03657955593532986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,12288,0.01889244384235806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,3072,0.016846223009957206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,8192,0.013714666995737286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,4096,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,768,0.017696888910399545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,256,0.01481599940194024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,256,0.015119999647140503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,2560,0.02165066699186961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,512,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,8192,0.22435466448465982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,7168,0.020249777370029025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,2048,0.008902221918106079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,256,0.002481777841846148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,1024,0.06017866399553087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,64,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,3584,0.011242667006121742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,768,0.017160000072585214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,3072,0.05867822302712334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,512,0.006911110960774952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,6144,0.040836443503697716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,768,0.03054311209254795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,10240,0.028370665179358587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,16384,0.021677333447668288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,128,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,32,0.0035128887328836652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,128,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,65536,0.01682400041156345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,8192,0.01071111112833023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,6144,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,512,0.0033973333322339584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,1536,0.0035111109415690103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,12288,0.01816177699300978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,65536,0.1329137749142117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,256,0.015484443969196744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,5120,0.00961599995692571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,512,0.014736000034544202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,65536,0.3017395602332221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,12288,0.03272799981964959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,32,0.00830844458606508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,768,0.14347288343641493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,64,0.04471910993258158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,768,0.020319110817379422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,1024,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,1024,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,5120,0.010122666756312052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,6144,0.026530666483773127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,128,0.015258666541841296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,1024,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,64,0.001839111041691568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,16384,0.43548001183403867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,4096,0.01997511088848114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,2048,0.005027555757098728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,1536,0.04814844330151876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,16384,0.02641333308484819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,3584,0.23385421435038248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,2560,0.007885333564546373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,5120,0.008967999782827165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,2560,0.004877333425813251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,6144,0.014820444915029736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,2048,0.0124195557501581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,128,0.0052764444715446895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,1536,0.01407999959256914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,128,0.009515555368529426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,1024,0.012383999923865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,3072,0.056074665652381055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,256,0.0028159999185138275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,8192,0.0203484445810318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,5120,0.008738666772842407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,256,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,512,0.07436888747745089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,2560,0.013983999689420065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,3072,0.013168888787428537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,256,0.0028560000161329904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,8192,0.019303111566437613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,5120,0.02384533319208357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,4096,0.009307555854320526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,3072,0.012653333445390066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,1536,0.070433775583903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,10240,0.02956533432006836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,2560,0.010905777414639791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,768,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,256,0.01647911138004727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,1024,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,256,0.015439111325475903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,1024,0.007308444215191736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,5120,0.01852799952030182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,12288,0.03739022215207418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,1024,0.00562755556570159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,256,0.01478488908873664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,3072,0.007311999797821045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,3072,0.02365511159102122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,256,0.01036088913679123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,16384,0.04733688963784111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,1536,0.05614755551020304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,256,0.00221422231859631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,10240,0.30346666442023384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,256,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,2048,0.02311822275320689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,512,0.051557332277297974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,3072,0.007335111498832703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,16384,0.02218311031659444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,4096,0.00795733349190818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,5120,0.010018666585286459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,12288,0.033434665865368314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,512,0.02495999965402815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,5120,0.11643200450473362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,7168,0.03782577647103204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,1024,0.003346666693687439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,128,0.0195448895295461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,2048,0.04674933354059855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,32,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,32,0.0032302221904198327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,8192,1.7624293433295355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,256,0.014764444695578681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,3584,0.019889778561062284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,32,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,2048,0.14370489120483398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,7168,0.03612533211708069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,16384,0.2640319930182563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,10240,0.019874667127927143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,1536,0.027789334456125896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,768,0.003969777789380816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,8192,0.013576000101036496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,768,0.003980444537268745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,64,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,2560,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,65536,0.09299555752012466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,65536,0.09832177559534709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,128,0.014482667048772177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,10240,0.016136889656384785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,10240,0.01274400038851632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,10240,0.06447911262512207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,16384,0.14760355154673258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,2560,0.017854221993022494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,8192,0.14441422621409097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,6144,0.02100533412562476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,32,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,1536,0.0069146665434042616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,3072,0.007988444632954067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,128,0.010378667049937779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,1536,0.01239733315176434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,1536,0.008633777499198914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,512,0.06012533108393351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,5120,0.016252444850073922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,2048,0.014410666293568082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,16384,0.027446221974160936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,3072,0.007271111011505127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,3072,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,3584,0.014807111687130399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,65536,0.47004175186157227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,8192,0.0125937776433097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,6144,0.016474665866957772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,5120,0.019137778215938143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,2048,0.02260444396071964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,5120,0.016514667206340365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,32,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,768,0.0028782222006056043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,3072,0.012379555238617791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,3072,0.007931555310885111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,2560,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,6144,0.00998933364947637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,7168,0.027021333575248718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,3072,0.019838222199016146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,6144,0.01720622181892395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,4096,0.025360888904995386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,3584,0.007598222129874759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,6144,0.022600000103314716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,3584,0.04145155681504144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,1024,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,65536,0.19269333945380318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,1536,0.004858666823969947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,65536,0.14778844515482584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,5120,0.009685333404276106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,4096,0.038104888465669416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,7168,0.010567111273606619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,2048,0.006220444622966979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,64,0.0031475555151700974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,2048,0.014431110686726041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,10240,0.028512001037597656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,5120,0.034635557068718806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,128,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,10240,0.027480888697836135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,12288,0.03939111034075419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,10240,0.019869332512219746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,32,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,3072,0.031217777066760596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,768,0.0057760001056724125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,1536,0.017491555876202054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,65536,1.9089590708414714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,4096,0.03439111179775662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,3584,0.007264888948864407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,65536,1.1605581707424588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,512,0.014755555325084262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,4096,0.045464889870749585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,8192,0.026649778087933857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,1024,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,16384,0.11645778020222981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,4096,0.11963733037312825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,128,0.012314666476514606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,4096,0.17792710993025038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,2560,0.01688799924320645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,768,0.006161777923504512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,10240,0.02366133365366194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,1536,0.011819555527634092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,5120,0.14895199404822454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,32,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,4096,0.008231999973456064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,768,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,3584,0.01571377780702379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,6144,0.009309333231714036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,4096,0.008106666306654612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,65536,0.1084773341814677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,5120,0.10929510990778606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,128,0.006310222049554189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,8192,0.005621333503060871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,1024,0.014926221635606555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,16384,0.017143110434214275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,2560,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,1536,0.011773332953453064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,1536,0.014925332532988654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,3072,0.02808977829085456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,768,0.017863111363516915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,2560,0.023041778140597876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,3584,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,2048,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,65536,0.08219377862082587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,7168,0.025781333446502686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,4096,0.011376889215575324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,3072,0.0124195557501581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,1024,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,16384,0.0229795558585061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,2048,0.008012444608741337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,10240,0.0465262234210968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,65536,0.18984533680809867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,8192,0.011024889018800525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,2048,0.00453333349691497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,12288,0.02369333306948344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,128,0.015096889601813423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,512,0.026717333330048457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,1024,0.0031999999450312722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,128,0.01482222146458096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,1536,0.006636444479227066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,1024,0.005416888743638992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,1536,0.03632444474432204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,7168,0.012727110750145383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,1024,0.003658666792843077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,6144,0.009343999955389235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,128,0.0018257778137922287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,128,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,1536,0.19868000348409018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,1536,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,65536,0.5490666495429145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,16384,0.031276444594065346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,3072,0.307640896903144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,5120,0.013398222625255585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,5120,0.07514311207665338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,768,0.004885333279768626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,2048,0.00831200016869439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,128,0.015824000040690105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,4096,0.010323555933104621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,1024,0.1291217803955078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,4096,0.07851288715998332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,1024,0.07242755757437812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,7168,0.020059555768966675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,12288,0.021553777986102637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,512,0.01143555591503779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,3584,0.007857778006129796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,768,0.00254844439526399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,256,0.025031111306614343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,65536,0.18192622396681044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,6144,0.009167110754383935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,1536,0.0038906666967603895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,2048,0.06252088811662462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,12288,0.02456888887617323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,1024,0.005964444329341252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,12288,0.01313333296113544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,2560,0.009563555320103964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,3072,0.012721777790122561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,3584,0.01166488892502255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,3584,0.016148444679048326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,12288,0.017120000388887193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,768,0.018790221876568265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,1536,0.01216533366176817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,768,0.017961778574519686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,1536,0.010328888893127441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,512,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,32,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,12288,0.2819484339820014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,32,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,65536,0.10022933615578546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,1024,0.006258666515350342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,128,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,7168,0.06601866748597887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,3072,0.01702577703528934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,3072,0.028012444575627644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,7168,0.019259555472267997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,4096,0.007600000335110559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,2560,0.010445333189434476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,7168,0.01407199932469262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,1536,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,10240,0.026699556244744196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,8192,0.023864888482623633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,8192,0.03330311179161072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,2560,0.01256444470749961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,128,0.009196444518036312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,6144,0.009348444640636444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,10240,0.04261599977811178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,3584,0.01834488908449809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,64,0.001867555495765474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,5120,0.01627111103799608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,128,0.005203555441564984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,8192,0.1027182208167182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,768,0.004047111090686587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,10240,0.03247111042340597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,768,0.047057777643203735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,1536,0.005199999858935674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,32,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,3584,0.007311999797821045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,6144,0.00980799976322386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,32,0.008431110945012834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,3072,0.013140444126394061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,1024,0.0028524444335036804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,7168,0.01960711181163788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,7168,0.007986666427718269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,8192,0.016015999846988253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,1536,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,2560,0.02327288852797614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,1024,0.012529777983824411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,10240,0.017997332745128207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,4096,0.02975911233160231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,128,0.01849422189924452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,65536,0.15762489371829563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,16384,0.13800266053941515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,7168,0.33730223443773055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,12288,0.016805332567956712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,10240,0.02241155505180359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,256,0.014774221513006421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,1536,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,5120,0.1432373391257392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,1024,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,3584,0.00792711145348019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,3072,0.007250666618347168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,1024,0.009687999884287516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,768,0.003620444486538569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,3584,0.02407377792729272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,6144,0.009382222261693742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,10240,0.017436444759368896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,7168,0.019561777512232464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,7168,0.010600888894663917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,1536,0.016301333904266357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,4096,0.022832888695928786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,12288,0.034048000971476235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,1024,0.007320889168315464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,4096,0.008279110822412703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,10240,0.6547182401021322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,3072,0.023444443941116333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,8192,0.2035271061791314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,7168,0.021987555755509272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,4096,0.01827911039193471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,4096,0.013826666606797112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,64,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,65536,0.25068532096015084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,768,0.008611555728647444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,32,0.0022053333620230355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,3584,0.01867555578549703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,7168,0.07099111212624444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,3072,0.015784000356992085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,3584,0.00792622235086229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,12288,0.02149244480662876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,10240,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,3584,0.013760888742076026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,4096,0.054660442802641124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,6144,0.01496266656451755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,256,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,4096,0.008274666965007782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,5120,0.030335999197430078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,7168,0.01276266657643848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,2048,0.052278220653533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,10240,0.005623999983072281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,4096,0.008556444611814287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,3584,0.04502844479348925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,6144,0.00904088881280687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,128,0.005944889038801193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,3072,0.03674755493799845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,5120,0.11867911285824245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,512,0.012784889174832238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,65536,1.3162221908569336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,12288,0.016908443636364408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,128,0.0058782220714622075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,3584,0.0041777777175108595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,256,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,256,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,3584,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,7168,0.010034666293197209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,5120,0.017709333035680983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,3072,0.011383111278216044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,4096,0.008676444490750631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,256,0.014801777071423002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,4096,0.03672711054484049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,10240,0.004879999905824661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,12288,0.02936089038848877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,1536,0.021264889174037512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,3584,0.01737866633468204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,65536,0.9957653681437174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,7168,0.014190221826235453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,16384,0.024707555770874023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,65536,0.012059555285506778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,4096,0.017255999975734286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,65536,0.4667217996385362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,512,0.010210666391584609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,6144,0.015474667151769003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,256,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,6144,0.04212266537878248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,7168,0.02388711108101739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,6144,0.0196426659822464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,1536,0.006268444574541516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,1536,0.0052639999323421055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,4096,0.012903110848532783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,512,0.0028204443968004654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,3584,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,5120,0.004239999999602635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,10240,0.0535768899652693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,64,0.009699555734793345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,5120,0.017695999807781644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,64,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,128,0.007267555428875818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,6144,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,64,0.0021315556433465746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,6144,0.018780445059140522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,256,0.006642666541867786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,1024,0.006575111299753189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,3584,0.010670222342014313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,1536,0.008198222352398766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,2048,0.00518844442235099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,1536,0.004200889004601372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,1024,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,5120,0.008665777742862701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,1536,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,128,0.0052044445441828836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,1536,0.012420444852775998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,1536,0.0606515540017022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,4096,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,16384,0.019190222024917603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,5120,0.012065778175989786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,3584,0.09414755635791355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,512,0.02113866640461816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,3072,0.007623111208279927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,64,0.0021235555824306277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,64,0.00794400026400884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,2560,0.006959111326270633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,3584,0.021985777550273474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,1536,0.007187555233637492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,7168,0.013389333254761167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,12288,0.07237155569924249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,3584,0.015814221567577787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,64,0.015522667103343539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,2560,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,2048,0.012051555845472546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,7168,0.016526222229003906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,3584,0.011800000237094032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,16384,0.03464533223046197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,5120,0.09809688727060954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,256,0.005207111024194294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,3584,0.007366221812036302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,2048,0.01991555591424306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,6144,0.02640177806218465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,7168,0.009932444327407414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,8192,0.011698666546079846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,1536,0.006223111102978389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,2560,0.02740266587999132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,64,0.006907555378145642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,1536,0.006638222270541721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,1024,0.003495111233658261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,10240,0.6900889078776041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,256,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,65536,0.1445102161831326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,12288,0.02103022237618764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,6144,0.01348888874053955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,6144,0.03379022081693014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,1536,0.017204445269372728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,65536,0.09663999742931789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,4096,0.01888177792231242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,32,0.0019884444773197174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,6144,0.018552889426549275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,512,0.01406311160988278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,2048,0.006923555499977536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,2560,0.020245333512624104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,512,0.02959111001756456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,5120,0.023753777146339417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,5120,0.024561777710914612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,5120,0.02532800038655599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,16384,0.010242666635248397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,8192,0.026218665970696345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,64,0.002536888958679305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,3584,0.021928888228204515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,5120,0.09367822276221381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,2048,0.006940444310506185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,6144,0.015519110692871941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,64,0.011391999820868174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,4096,0.017671111557218764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,7168,0.02793244520823161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,64,0.002163555473089218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,65536,0.1691208945380317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,10240,0.0336808893415663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,2560,0.04993066522810194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,512,0.01753866672515869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,8192,0.2994115617540148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,65536,0.10831999778747559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,256,0.018032888571421306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,12288,0.021128000484572515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,5120,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,5120,0.0107004443804423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,64,0.0025155554629034465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,256,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,3584,0.004912000149488449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,3072,0.013895110951529609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,1536,0.031425777408811785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,768,0.008721777962313758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,768,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,1024,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,1024,0.0031902222997612427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,4096,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,768,0.05209333366817898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,2560,0.18048177825080025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,4096,0.02367377777894338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,3072,0.004561777744028303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,6144,0.020021332634819877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,5120,0.03541510966089036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,768,0.013448889056841532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,5120,0.008638222184446123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,256,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,5120,0.018599111172888014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,2048,0.004568888909286923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,256,0.0022071111533376905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,3072,0.005533333453867171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,6144,0.11661067273881699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,256,0.02932000160217285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,4096,0.015788444214397006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,2560,0.006909333169460297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,128,0.0058702222175068325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,65536,0.06746221913231744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,1024,0.003499555504984326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,16384,0.03347200155258179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,5120,0.016183111402723525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,128,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,3072,0.02181688944498698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,7168,0.446549309624566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,16384,0.01956888867749108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,5120,0.02477688921822442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,1536,0.020796444680955675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,16384,0.024902222885025874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,4096,0.015126221709781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,512,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,1536,0.023987554841571387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,2560,0.028824888997607764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,3072,0.009584888815879822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,8192,0.0557253360748291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,32,0.007248889240953658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,8192,0.012536889149083031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,10240,4.944431304931641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,7168,0.01096088853147295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,256,0.006680888848172293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,128,0.015475556254386902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,768,0.020695999264717102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,2048,0.0322675539387597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,8192,0.012503999802801343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,1536,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,256,0.010729778144094678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,8192,0.022711111439598933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,1024,0.006329777754015393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,1024,0.007327111230956183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,512,0.005299555758635203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,5120,0.02110933264096578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,4096,0.011049778097205691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,8192,0.04529244369930691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,512,0.01752622259987725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,65536,0.2992844316694471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,16384,0.08202844195895724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,65536,0.08057777749167548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,65536,0.14848000473446316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,3072,0.007293333609898885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,16384,0.020626667473051283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,16384,0.07182844479878743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,64,0.0021688888470331826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,16384,0.020304888486862183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,8192,0.014064888159434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,64,0.006743999818960826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,6144,0.017435555656750996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,2048,0.007594666547245449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,7168,0.044616887966791786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,4096,0.3204977777269152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,3584,0.012344888514942594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,10240,0.014321777555677625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,256,0.01551555593808492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,256,0.002489777695801523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,64,0.006597333484225803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,3072,0.013778666655222574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,3584,0.01038755559258991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,1024,0.0028284444577164124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,8192,0.06403822369045682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,1024,0.00565244464410676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,10240,0.04438844323158264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,8192,0.049935112396876015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,4096,0.0079297779334916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,2560,0.019283556275897555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,65536,0.06615822182761298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,3584,0.004909333255555895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,4096,0.012708444562223224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,2560,0.007270221908887227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,5120,0.19204355610741508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,512,0.00282577777074443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,5120,0.011737777955002256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,12288,0.014306666122542487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,128,0.0148044443792767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,1536,0.07488266626993816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,2560,0.012930666406949362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,64,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,768,0.01686755485004849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,7168,0.009309333231714036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,3584,0.03488622109095255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,768,0.006950221955776215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,128,0.0038604442444112566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,8192,0.04525599877039591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,1536,0.053730666637420654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,16384,0.05206222335497538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,128,0.012418666647540199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,2560,0.009208888643317753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,12288,0.1578871144188775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,2560,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,768,0.019339554839664035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,1536,0.015116444892353482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,64,0.00758133331934611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,5120,0.007816889219813878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,32,0.03680888811747233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,3072,0.014172444740931192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,12288,0.018212444252438016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,2048,0.011233777635627322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,4096,0.0247653325398763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,1024,0.006116444451941385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,1536,0.012438221938080259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,768,0.03016177813212077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,128,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,65536,0.15231200059254965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,2048,0.0069413334131240845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,768,0.002846222163902389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,8192,0.09557777643203735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,1536,0.0038284444146686127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,768,0.005554666535721884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,256,0.005201777650250329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,128,0.014477334088749357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,8192,0.019905777441130746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,2048,0.006635555376609166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,1024,0.015252444479200574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,512,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,1024,0.0031715554909573663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,2560,0.006923555499977536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,512,0.014130666851997375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,3584,0.08146844307581584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,65536,1.5801742341783311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,128,0.017438222964604694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,4096,0.012103999654452005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,12288,0.043078223864237465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,64,0.011708444191349877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,1024,0.009293333523803288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,1536,0.003963555726740095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,32,0.005640888793600931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,8192,0.012867555850081973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,65536,0.04723822077115377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,1536,0.00654755574133661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,1024,0.018955555227067735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,1536,0.012438221938080259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,256,0.002477333363559511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,65536,0.08895911110772027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,16384,0.03968355390760634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,1536,0.007637333538797166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,2048,0.11531021859910752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,3072,0.012789333032237159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,65536,0.13668089442782932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,16384,0.03747377792994181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,128,0.005236444373925527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,1536,0.019766221443812054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,8192,0.041768001185523145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,1024,0.012097777591811286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,6144,0.01330933305952284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,128,0.022111111217074927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,16384,0.15060178438822427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,32,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,6144,0.06681955522961087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,2048,0.029080000188615587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,1024,0.00629333323902554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,2048,0.013086222112178802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,512,0.0028088889602157804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,512,0.0032968889507982465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,1024,0.006182222316662471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,10240,0.02602400051222907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,65536,0.06328888734181722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,8192,0.024480889240900677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,10240,0.02289955483542548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,5120,0.01646577815214793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,1024,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,2560,0.05952533086140951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,4096,0.17625778251224092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,128,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,10240,0.018238221605618794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,128,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,2048,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,1024,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,6144,0.06586044364505343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,32,0.015135111080275642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,512,0.014102222190962898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,256,0.0053635554181204895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,5120,0.012819555898507437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,1024,0.04408444298638237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,5120,0.0296088887585534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,8192,0.019606222709019978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,4096,0.04307377669546339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,2048,0.008637333081828224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,65536,0.055142223834991455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,32,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,512,0.005326222214433882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,3584,0.01092888911565145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,2560,0.01609688831700219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,512,0.00905688852071762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,8192,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,2560,0.00867555538813273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,7168,0.0214373336897956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,4096,0.015460444821251763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,2560,0.05316533313857185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,8192,0.014840889308187695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,6144,0.0436062216758728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,1536,0.07659288909700182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,1024,0.0199297782447603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,2560,0.020059555768966675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,8192,0.038535998927222356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,2048,0.027952889601389568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,3584,0.02358666724628872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,10240,0.015985776980717976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,4096,0.020272889071040683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,12288,0.011535111400816174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,1024,0.011695110963450538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,512,0.015166223049163818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,2048,0.004618666652176115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,768,0.01886044442653656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,7168,0.00831733312871721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,16384,0.012914666699038612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,128,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,5120,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,256,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,12288,0.022820444570647344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,128,0.039388444688585066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,2560,0.02196088929971059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,2560,0.007078222102589077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,3584,0.006621333460013072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,4096,0.032631112469567194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,1024,0.013097777962684631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,128,0.0025048889219760895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,2560,0.004600000050332812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,10240,0.014333332578341166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,65536,0.14841332700517443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,5120,0.011178666518794166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,6144,0.01203733351495531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,256,0.010017777482668558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,256,0.12058044804467095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,8192,0.011981333295504252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,1024,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,256,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,3584,0.03348799877696567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,256,0.024404444628291663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,2048,0.008406221866607666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,3072,0.00830577810605367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,3072,0.007326222128338284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,512,0.0031848889258172777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,4096,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,2560,0.028704888290829126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,7168,0.6238417625427246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,6144,0.013425778183672162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,16384,0.29973867204454213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,4096,0.022389334109094407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,512,0.0035297779573334586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,1536,0.008562666674455007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,7168,0.012715555727481842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,2048,0.02003466586271922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,2560,0.007986666427718269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,10240,0.041914665036731295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,12288,0.40277332729763454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,32,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,12288,0.016143111719025504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,512,0.007617777420414819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,7168,0.0668391121758355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,12288,0.13434399498833552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,10240,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,2560,0.012345777617560493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,2560,0.014121777481502958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,1536,0.008759111166000366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,3584,0.009019555317031013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,2560,0.024354666471481323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,512,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,12288,0.08083733585145739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,3072,0.009319999979601966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,4096,0.02230311102337307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,768,0.015134221977657743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,512,0.005407111098368962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,256,0.01108977778090371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,1536,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,256,0.014085332552591959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,16384,0.030249777767393324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,64,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,1024,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,12288,0.06652355856365628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,256,0.1432960033416748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,768,0.012741333080662621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,1536,0.00444088876247406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,2560,0.07384000221888225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,65536,2.4399049546983504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,8192,0.021504888931910198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,2560,0.04408266809251574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,7168,0.17106578085157606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,8192,0.06447199980417888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,6144,0.391590224372016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,10240,0.026611554953787062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,16384,0.01646399994691213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,6144,0.0075911109646161394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,128,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,6144,0.016022221909628976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,2048,0.013188444077968597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,256,0.013772444592581855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,3072,0.02199733257293701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,2560,0.03126488791571723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,1024,0.007608888877762689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,64,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,16384,0.04799377918243408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,3072,0.0076453329788313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,1536,0.010006222460005019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,3072,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,65536,0.16899377769894072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,16384,0.023640000157886084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,16384,0.019208000765906442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,10240,0.016421332955360413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,256,0.011409777734014722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,1024,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,65536,0.09753066963619655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,65536,0.06282755401399401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,10240,0.016181333197487723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,65536,0.5399493111504449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,512,0.025417778227064345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,1536,0.00811111099190182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,3072,0.017498667041460674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,4096,0.008943999807039896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,512,0.04473155405786303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,10240,0.007617777420414819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,6144,0.015136889285511441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,64,0.002151111140847206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,2048,0.013809777796268463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,3584,0.025080889463424683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,10240,0.07644355297088623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,12288,0.09702133470111424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,3072,0.018583110637134977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,1536,0.0099653336736891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,12288,0.01000088867213991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,512,0.013787555197874704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,768,0.00315022220214208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,64,0.019015999303923715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,65536,0.06605066855748494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,6144,0.024989333417680528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,4096,0.05109066764513651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,12288,0.040581332312689886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,3584,0.008303110798199972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,5120,0.022264889544910852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,64,0.0058764442801475525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,2560,0.008280889027648503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,4096,0.032383110788133405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,6144,0.01538044379817115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,4096,0.009129777550697327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,65536,0.05681066380606758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,12288,2.6349991692437067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,5120,0.014279999666743808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,1024,0.0035617777870761026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,8192,0.011060444845093621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,4096,0.01014666673209932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,10240,0.016511999898486666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,1536,0.01182044463025199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,3072,0.008300444318188561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,12288,0.023356444305843774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,10240,0.004510222209824456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,128,0.0038862224254343244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,32,0.0028595555987623003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,4096,0.008564444051848518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,512,0.006847111301289664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,1536,0.013430222041077085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,4096,0.017253332667880587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,65536,0.19272177749209932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,8192,0.015249777171346875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,65536,0.7393137613932291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,3072,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,128,0.0026400000270869997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,4096,0.011687111523416309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,512,0.003157333367400699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,2048,0.018203554881943595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,8192,0.01276711126168569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,3072,0.007618666523032718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,256,0.006321777900060018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,16384,0.021643555826610986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,128,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,2560,0.0195448895295461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,5120,0.01719555589887831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,65536,0.075126224093967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,5120,0.02719466719362471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,128,0.0021182222084866627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,8192,0.04037955403327942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,3584,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,4096,0.008374222450786168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,3584,0.004863111095296012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,4096,0.007972444097201029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,2048,0.014132445057233175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,2560,0.0211191111140781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,8192,0.010703999963071613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,12288,0.10142666763729519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,256,0.012223111258612739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,32,0.002536888958679305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,3072,0.017345777816242643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,16384,0.030324445830451116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,64,0.03346844514211019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,16384,0.57750489976671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,5120,0.018228444788191054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,16384,0.018559111489189994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,5120,0.004889777965015835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,1536,0.01887911061445872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,512,0.0031804444475306403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,7168,0.023504889673656885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,1024,0.012749333348539142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,5120,0.5400906668768989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,256,0.0021386666016446217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,128,0.014811555544535318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,6144,0.059432890680101186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,768,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,256,0.0024604443460702896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,512,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,1024,0.016194666425387066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,4096,0.009165333376990424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,10240,0.017864889568752713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,12288,0.03030844529469808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,2048,0.014136888914638095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,16384,0.043848888741599194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,6144,0.018964444597562153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,8192,0.005207111024194294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,64,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,16384,0.02423288921515147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,1024,0.06704711251788668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,1536,0.007927999728255803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,5120,0.01841600073708428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,256,0.00519822206762102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,65536,0.055382221937179565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,1024,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,12288,0.03632622294955783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,4096,0.09348089165157741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,64,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,12288,0.019927110936906602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,2048,0.008813333180215623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,32,0.003415999934077263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,3072,0.011375111010339526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,5120,0.013433777623706393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,5120,0.007664889097213745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,4096,0.009034666750166152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,32,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,3584,0.08372444576687282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,4096,0.01849688920709822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,256,0.007633777956167857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,16384,0.021896888812383015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,2048,0.013817778064144982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,2560,0.00847999999920527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,16384,0.37042755550808376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,1536,0.017128000656763714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,1024,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,10240,0.03394488824738397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,5120,0.02645955483118693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,3072,0.01109688894616233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,512,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,256,0.016576889488432143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,8192,0.014287999934620328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,64,0.007655999726719326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,256,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,2560,0.02234755622016059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,3584,0.011708444191349877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,128,0.011388444238238864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,64,0.0020666666742828158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,768,0.005961777849329843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,8192,0.022608889473809138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,12288,0.014473777678277759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,768,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,256,0.0024604443460702896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,10240,0.02459377712673611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,2048,0.010010666317409938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,32,0.005610666755172942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,1536,0.0058746664888328975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,2048,0.004545777622196409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,4096,0.008994667066468133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,1024,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,256,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,3072,0.007640000018808577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,12288,0.02371999952528212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,768,0.006268444574541516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,10240,0.02870133188035753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,128,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,3072,0.05956177579032051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,16384,0.029674665795432195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,768,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,12288,0.01680266691578759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,768,0.01867822143766615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,2560,0.011490667031870948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,3072,0.007860444486141205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,1024,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,512,0.006637333167923822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,1536,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,1536,0.011043555206722684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,3584,0.025701334079106648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,2560,0.015415110521846347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,1536,0.013420444395807056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,5120,0.1502702236175537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,6144,0.010077333284748925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,768,0.002855111120475663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,128,0.0018515555808941524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,12288,0.010440888504187265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,65536,0.08487644460466172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,256,0.006593777901596493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,7168,0.21530400382147896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,3072,0.008655110994974772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,12288,0.014420444766680399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,16384,0.03804533349143134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,32,0.008292444050312042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,65536,0.057465778456793896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,5120,0.013292444248994192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,32,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,6144,0.05624000231424967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,1536,0.009312888814343346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,12288,0.02016088863213857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,1536,0.013453333742088742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,3584,0.008271111382378472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,1536,0.008285333712895712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,5120,0.034246222840415105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,8192,0.16172266006469727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,32,0.0021626665774318907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,16384,0.03790400094456143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,1536,0.003540444291300244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,2048,0.018292443619834054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,12288,0.7549333572387695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,12288,0.08079999685287476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,8192,0.12953866852654353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,2560,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,3584,0.0217866665787167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,6144,0.2744657728407118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,3072,0.016396444704797532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,128,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,7168,0.01275466630856196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,32,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,128,0.008303110798199972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,3072,0.005012444323963589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,6144,0.025058666865030926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,3584,0.013349333571063148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,64,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,4096,0.015788444214397006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,7168,0.010107555323176915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,10240,0.011705777711338468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,3072,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,8192,0.015433778365453085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,10240,0.011733333269755045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,1024,0.014763555592960782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,16384,0.23408177163865831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,2560,0.012907555533779992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,6144,0.011023110813564725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,64,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,2048,0.023741333021057978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,6144,0.17007111178504097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,5120,0.019428445233239066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,1536,0.007958222594526079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,8192,0.0158897770775689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,2560,0.0420702232254876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,3584,0.008039111064540016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,256,0.015098666151364645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,768,0.007280000381999546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,7168,0.03070488903257582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,6144,0.035855998595555626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,16384,0.0341635545094808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,3584,0.04311911265055338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,8192,0.01069155583779017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,1024,0.002850666642189026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,8192,0.032808889945348106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,2048,0.013493333425786761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,64,0.0028497777465316984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,65536,0.018263111511866253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,1536,0.015823110938072205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,6144,0.028821332587136164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,64,0.008957333034939235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,3584,0.017885333961910672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,3584,0.01763199932045407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,512,0.005640000104904175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,64,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,256,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,128,0.0063777777055899305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,32,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,1024,0.02278933260175917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,32,0.011367999845080905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,512,0.011060444845093621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,65536,0.1386151048872206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,128,0.006593777901596493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,10240,0.01754844519827101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,10240,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,3584,0.07163111368815105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,1024,0.05235377616352505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,2560,0.08710222111807929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,2560,0.03341066506173875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,2048,0.01511111193233066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,3584,0.05613778034845988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,256,0.005209777918126848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,512,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,2048,0.0069324444565508105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,6144,0.027196443743175928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,6144,0.009349333743254343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,3072,0.020265777905782063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,1536,0.009029332962301042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,3072,0.017872000734011333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,128,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,3072,0.011129777464601727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,12288,0.016511110795868766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,65536,4.451134999593099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,6144,0.008981333010726506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,32,0.007986666427718269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,12288,0.01714666684468587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,3584,0.014871999621391296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,2560,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,7168,0.026716444227430556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,128,0.005395555661784277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,8192,0.008423111504978603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,12288,0.017229333519935608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,32,0.003879111260175705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,128,0.002838222309947014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,8192,0.010814222196737925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,128,0.01495733360449473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,16384,0.051035556528303355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,4096,0.07041866911782159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,8192,0.0915715561972724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,6144,0.010024000373151567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,1024,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,7168,0.012400888734393649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,256,0.002842666581273079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,10240,0.01309511148267322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,3072,0.010838222172525195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,2560,0.01685244507259793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,16384,0.09381333324644302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,768,0.008233778178691864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,4096,0.049618668026394315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,12288,0.05323555403285556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,256,0.005695999910434087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,2048,0.016142222616407607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,8192,0.02346044447686937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,5120,0.011047999891969891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,12288,0.04158844550450643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,1536,0.006171555568774541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,2560,0.009663999908500248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,7168,0.031174222628275555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,256,0.0196142229768965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,768,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,3072,0.037393778562545776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,2560,0.008544888761308458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,16384,0.020949333906173706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,128,0.015495111544926962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,128,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,4096,0.016929777132140267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,16384,0.07850933074951172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,6144,0.0123857781291008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,1024,0.09193155500623916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,8192,0.011756444142924415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,4096,0.0928106639120314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,8192,0.0236817780468199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,16384,0.022387555903858606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,256,0.016813332835833233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,768,0.015462223026487561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,12288,0.06741066773732503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,7168,0.03332355618476868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,8192,0.03562755717171563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,65536,1.1463777754041884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,2560,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,128,0.014464888307783337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,10240,0.013435555828942193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,256,0.01478488908873664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,512,0.007744889292452071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,10240,0.02584533393383026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,128,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,512,0.007244444555706448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,512,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,512,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,768,0.0034702221552530923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,1024,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,4096,0.016196444630622864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,3584,0.021201777789327834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,2560,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,768,0.0032026666320032547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,1536,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,7168,0.12168799506293403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,3584,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,8192,0.025391111771265667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,3072,0.009025777379671732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,1536,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,8192,0.06100710895326403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,128,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,16384,0.048002666897243924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,8192,0.013065777719020844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,6144,0.03442400031619602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,65536,0.06313244501749675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,7168,0.029280887709723577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,1536,0.0148044443792767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,768,0.017557332913080852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,3072,0.007242666350470648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,10240,0.0177831103404363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,2560,0.04592355423503452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,3584,0.04182044333881802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,6144,0.013015111287434896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,2560,0.007650666766696506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,512,0.013069333301650153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,2560,0.010274666878912184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,7168,0.01512977812025282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,10240,0.2721964518229167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,2048,0.008330666356616551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,768,0.006226666685607698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,7168,0.11344444751739502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,128,0.014878221684032016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,2560,0.008958222137557136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,3584,0.008273777862389883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,10240,0.03626577721701728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,256,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,3584,0.025735111700163946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,12288,0.03151111139191522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,7168,0.17726043860117593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,3072,0.007312888900438945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,6144,0.012479110724396177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,1024,0.012344888514942594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,3584,0.0698453320397271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,10240,0.02455288834042019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,128,0.007723555796676212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,8192,0.022290666898091633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,3584,0.01484266585773892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,512,0.017275555266274344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,7168,0.2937297821044922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,512,0.002472888885272874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,8192,0.008107555409272512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,768,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,2560,0.11057777537239923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,3584,0.03209777673085531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,12288,0.04106311003367106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,512,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,256,0.005894222193294101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,768,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,512,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,2048,0.07896711428960164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,256,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,16384,0.017907554904619854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,16384,0.03719377848837111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,65536,0.2896764543321398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,4096,0.008968888885445064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,3072,0.013959111438857185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,1024,0.12137333552042644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,1024,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,2560,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,2048,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,2560,0.012120000190205045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,512,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,7168,0.03863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,256,0.020265777905782063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,16384,0.4270302189721002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,16384,1.113290680779351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,65536,0.19073688983917236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,6144,0.027091556125217017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,768,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,768,0.14620799488491482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,3584,0.02719377809100681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,2560,0.008988444175985124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,256,0.014782221780882942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,1024,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,768,0.0210817770825492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,3072,0.01482399966981676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,16384,0.14055111673143175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,32,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,4096,0.22097955809699166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,16384,0.05729689200719198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,6144,0.008991111483838823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,1024,0.02390755547417535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,2560,0.015659555792808533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,7168,0.010326222413116032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,64,0.006605333338181178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,5120,0.004590222405062782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,16384,0.5391768879360622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,7168,0.08063555426067777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,16384,0.021196444829305012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,8192,0.07267822159661187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,128,0.006907555378145642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,32,0.013759111364682516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,64,0.01038755559258991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,128,0.032079110542933144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,65536,0.06367555591795179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,12288,0.023992889457278784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,4096,0.036869333850012884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,768,0.004224888980388641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,1024,0.0075288886825243635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,4096,0.03579644362131754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,16384,0.026760889424218073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,256,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,8192,0.015559110376569959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,3072,0.009012444151772393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,128,0.0052675555149714155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,3584,0.007370666497283512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,2560,0.010326222413116032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,768,0.016559999850061204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,8192,0.014549333188268872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,128,0.014129777749379476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,65536,0.10266133149464925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,512,0.009354666703277165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,3584,0.024714666936132643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,32,0.09877600272496541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,5120,0.04509333438343472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,1536,0.02273688879277971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,12288,0.06610577636294894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,10240,0.030376000536812678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,65536,0.04782311121622721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,12288,0.1014640000131395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,2560,0.02460799945725335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,6144,0.009379555781682333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,6144,0.029483556747436523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,512,0.004536889079544279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,2048,0.004551110996140374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,65536,0.10002133581373428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,16384,0.6793200174967448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,6144,0.1416906648212009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,6144,0.0309840010272132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,5120,0.027457776996824477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,3584,0.009925333162148794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,6144,0.009361777868535783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,128,0.006620444357395172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,16384,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,6144,0.03363733490308126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,4096,0.020959110723601446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,3584,0.0136817776494556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,6144,0.011761777930789523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,6144,0.009334222310119206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,5120,0.02734666731622484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,32,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,1024,0.031888888941870794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,7168,0.028549333413441975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,64,0.004482666651407878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,5120,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,512,0.017982222967677645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,8192,0.023749333288934495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,3072,0.01572355628013611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,256,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,128,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,10240,0.014832889040311178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,10240,0.015432889262835184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,3584,0.010684444672531553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,5120,0.32688265376620823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,1536,0.09664533535639445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,3072,0.02477955487039354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,6144,0.013376889129479727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,1024,0.0032257777121331957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,4096,0.00867022242810991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,12288,0.062881780995263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,3584,0.022320000661744013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,2048,0.011219555305110084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,16384,0.25765689214070636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,65536,0.07675821913613214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,7168,0.013757333159446716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,128,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,32,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,16384,0.034133334954579674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,10240,0.012091555529170565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,1024,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,8192,0.42861154344346786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,64,0.00327822214199437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,10240,0.017976888351970248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,10240,0.03020533257060581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,12288,0.07704622215694852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,2048,0.00962400022480223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,8192,0.145179549853007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,1024,0.007654222349325816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,16384,0.02231377859910329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,256,0.005639111002286275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,7168,0.01353422221210268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,8192,0.010068444742096795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,5120,0.03298755486806234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,16384,0.045181334018707275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,3584,0.03737688726849026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,12288,0.11226844787597656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,2048,0.019309333629078336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,3584,0.016188444362746347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,3584,0.09030133485794067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,128,0.014155555102560254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,2560,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,256,0.0021191111041439907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,4096,0.01568622224860721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,32,0.0021253333737452826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,16384,0.021818665994538203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,768,0.018173333671357896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,12288,0.01444533301724328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,2048,0.016878222425778706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,2560,0.00887288898229599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,5120,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,5120,0.009760888914267222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,768,0.03143466512362162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,128,0.014450665977266101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,6144,0.02230666743384467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,7168,0.020819556381967332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,2048,0.016768000192112394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,768,0.017483555608325534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,7168,0.027235555979940627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,32,0.05525866813129849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,1024,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,7168,0.007272889216740926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,10240,0.009364444348547194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,1024,0.00453955555955569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,16384,0.03184533450338576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,1536,0.024043555061022442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,65536,0.21193333466847739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,2560,0.007072889142566257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,2560,0.04101422097947862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,10240,0.028093334701326158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,6144,0.013407111167907715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,8192,0.010729778144094678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,128,0.005218666460778978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,6144,0.018340445227093168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,1536,0.006224000205596288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,256,0.006288888967699475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,3584,0.06339555316501193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,768,0.005601777798599667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,1536,0.029426667425367568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,128,0.021864889396561518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,7168,0.012452444268597497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,16384,0.016172443826993305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,6144,0.020597333709398907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,256,0.014793778459231058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,12288,0.014850666125615438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,3584,0.0094355551732911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,5120,0.008351999852392409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,2560,0.011680888632933298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,65536,0.22750756475660536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,12288,0.06963466935687594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,128,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,7168,0.004557333472702238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,128,0.006304889089531369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,128,0.005676444619894028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,10240,0.02914311157332526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,5120,0.024870221813519795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,2048,0.016356445021099515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,256,0.005200888961553574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,768,0.01683822274208069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,256,0.006534222099516127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,16384,0.03128800127241347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,3072,0.022337777747048274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,4096,0.020873778396182593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,12288,0.5180880228678385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,1536,0.06710311439302233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,512,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,3584,0.08614666594399346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,512,0.00756177802880605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,2048,0.029410666889614526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,512,0.045702222320768565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,8192,0.02993333339691162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,2048,0.029693331983354356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,128,0.008664000365469191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,3072,0.21677333778805205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,768,0.004933333231343163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,2048,0.013861333330472311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,3072,0.004229333251714706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,128,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,1536,0.051903112067116626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,512,0.005953777581453323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,512,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,2560,0.005048888838953442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,3584,0.011056889262464313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,5120,0.03463911016782125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,6144,0.02957688768704732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,128,0.02555111050605774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,16384,0.0214720004134708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,16384,0.055071999629338585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,8192,0.040143112341562905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,128,0.006269333263238271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,1024,0.00795911086930169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,128,0.011418666276666852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,3584,0.008992888861232335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,2560,0.007154666715198093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,5120,0.026868444350030687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,3072,0.010667555862002902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,8192,0.010944888823562197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,12288,0.013783999615245394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,256,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,3584,0.03356000118785434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,7168,0.027058665951093037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,16384,0.39968710475497776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,12288,0.030679110023710463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,10240,0.0181048893266254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,2048,0.03627644313706292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,2048,0.019419555862744648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,65536,0.13109778033362493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,4096,0.02419288953145345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,16384,0.02777333392037286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,32,0.005976888868543837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,6144,0.03297777639495002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,768,0.03107644451989068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,10240,0.01605333387851715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,3072,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,8192,0.037646220790015325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,5120,0.0107360002067354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,2560,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,3072,0.03129333257675171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,768,0.009665778113736046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,64,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,16384,0.01348888874053955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,2560,0.029339555237028334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,2048,0.008692444198661381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,2048,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,2048,0.007280000381999546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,16384,0.46469688415527344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,768,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,65536,0.06700622373157077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,12288,0.018179555733998615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,5120,0.11366221639845107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,12288,0.013778666655222574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,128,0.005006222261322869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,1024,0.008008889026112026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,3584,0.021274665991465252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,1024,0.007864000068770515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,6144,0.26761778195699054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,5120,0.019738665885395475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,512,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,12288,0.02962933315171136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,128,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,3072,0.006829333388143116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,65536,0.20858932865990532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,3584,0.008026666939258575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,7168,0.004559111264016894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,128,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,1536,0.007293333609898885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,2560,0.006992888947327931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,5120,0.008976000050703684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,10240,0.011747555600272285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,512,0.0035439998739295532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,10240,0.011714666253990598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,1536,0.019807111885812547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,3584,0.01831288966867659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,1024,0.0031804444475306403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,12288,0.028303109937243994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,2560,0.013415999710559845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,128,0.013821333646774292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,2048,0.08094222015804715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,1024,0.012730666332774691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,1536,0.008655110994974772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,12288,0.14012711577945286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,1024,0.014781332678265043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,3584,0.010544000400437249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,3072,0.007282666862010956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,7168,0.0232640008131663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,128,0.006721777634488211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,128,0.008996444443861643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,768,0.023965333898862202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,7168,0.013078221844302284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,3584,0.007619555625650618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,64,0.003154666680428717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,1024,0.010869333313571082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,3072,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,16384,0.0786657796965705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,6144,0.012817777693271637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,64,0.009586667021115622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,4096,0.015597333510716757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,1024,0.016220443778567843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,256,0.005528888768619961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,256,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,512,0.010703110860453712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,6144,0.11894667148590088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,10240,4.0466736687554254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,128,0.020617778102556866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,10240,0.03138844503296746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,2048,0.004519999855094486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,768,0.006284444282452266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,10240,0.03636088967323303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,1536,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,8192,0.015806222955385845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,3584,0.007639110916190677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,6144,0.03171733352873061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,256,0.0021724444296624926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,10240,0.5341155793931749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,768,0.014859555496109856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,4096,0.01240888900227017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,10240,0.01274311128589842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,6144,0.03458044595188565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,7168,0.004945777770545748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,65536,0.10182399882210626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,6144,0.012182222472296821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,4096,0.025770665870772466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,256,0.010552888943089379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,1536,0.02533422244919671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,12288,0.023412444525294836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,65536,0.17785333262549508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,32,0.0018444444156355327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,1536,0.018851555056042142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,2048,0.008998221821255153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,1536,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,768,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,4096,0.027758222487237718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,1536,0.006604444649484422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,2560,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,2560,0.004916444420814514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,256,0.002507555608948072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,1536,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,1536,0.03926666577657064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,5120,0.04046400056944953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,10240,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,1024,0.006081777728266186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,5120,0.008871999879678091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,3072,0.020472000042597454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,8192,0.03915999995337592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,1024,0.017237333787812125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,512,0.0042133331298828125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,512,0.0028248888750871024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,1536,0.006226666685607698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,64,0.0018373332503769135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,16384,0.1049448913998074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,512,0.01514400045077006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,3584,0.015793777174419828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,12288,0.01754577789041731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,1536,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,64,0.0017911110901170308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,2048,0.025496888491842482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,4096,0.02446844511561924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,16384,0.02030044462945726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,4096,0.0158924443854226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,256,0.003506666670242945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,65536,0.06808000140719943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,3584,0.025962667332755193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,256,0.014459555347760519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,4096,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,2560,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,64,0.006255110932721033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,5120,0.00833333366447025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,8192,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,16384,0.02236088944805993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,65536,0.0638986693488227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,1536,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,1536,0.0034808889031410217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,3584,0.016753777861595154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,32,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,6144,0.029311110575993855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,1024,0.019142222073343065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,128,0.0052595556610160405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,7168,0.02829866607983907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,1536,0.08612977796130711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,65536,2.452231936984592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,16384,0.018013333280881245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,2048,0.008204444415039485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,64,0.0024888888001441956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,5120,0.02732622292306688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,2560,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,2560,0.008650666309727563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,10240,0.015169777803950839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,10240,0.5581333372328017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,2048,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,128,0.013417777915795645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,8192,0.03958933220969306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,512,0.0033004445334275565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,1024,0.007636444436179266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,8192,0.02756533357832167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,1024,0.009873777627944946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,7168,0.03048266636000739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,2560,0.03358488943841722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,16384,0.09402400255203247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,4096,0.00831288927131229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,32,0.002114666625857353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,1024,0.003168888803985384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,8192,0.01954844428433312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,3584,0.007745777567227681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,8192,0.015562666787041558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,1536,0.03879022267129686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,12288,0.04689955711364746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,512,0.006960888703664144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,4096,0.03895911243226793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,3584,0.024323556158277724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,1024,0.013745778136783175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,1024,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,768,0.005616888817813661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,512,0.0028773333049482773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,2048,0.012701333396964602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,8192,0.025622222158643935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,10240,0.017044444878896076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,2560,0.027789334456125896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,2560,0.012171555724408893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,7168,0.0329093337059021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,4096,0.04158044523662991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,1024,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,6144,0.01847022275129954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,256,0.005611555443869696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,65536,0.06032088730070326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,128,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,16384,0.09098755651050144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,3072,0.007820444802443186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,5120,0.01923288901646932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,128,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,768,0.01681600014368693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,128,0.01648888819747501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,12288,0.0321688883834415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,768,0.01615111033121745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,8192,0.15313422679901123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,128,0.014076444837782117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,2048,0.21573777993520102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,3584,0.01253600004646513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,12288,0.02460533380508423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,1536,0.00656088896923595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,16384,0.05121244324578179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,128,0.01958577831586202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,8192,0.04281066523657905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,7168,0.01240355521440506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,512,0.014078223043017916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,1024,0.01682133310370975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,3584,0.013401777380042605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,2048,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,256,0.01814844376511044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,16384,0.04242222176657783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,7168,0.04724977744950188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,2048,0.01275288893116845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,4096,0.00793688909875022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,512,0.0031600000543726813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,10240,0.012076444096035428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,7168,0.09506844149695502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,768,0.00424888895617591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,2560,0.01615111033121745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,10240,0.0166284441947937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,5120,0.013976888524161445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,10240,0.013959111438857185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,8192,0.1580924457973904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,768,0.02467022173934513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,32,0.00564177789621883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,10240,0.02595022155178918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,768,0.014874666929244995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,16384,0.04552711049715678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,32,0.001971555459830496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,2560,0.02671555512481266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,65536,0.7349075741238065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,128,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,7168,0.03264710969395108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,5120,0.014750222365061441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,4096,0.011943110989199745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,1024,0.003568888952334722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,4096,0.017487999465730455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,16384,0.06458755334218343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,2560,0.00829777783817715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,64,0.008984000318580205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,3584,0.014449778530332776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,256,0.006305777778228124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,2560,0.13054933812883165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,2560,0.014856888188256158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,256,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,8192,0.04368889000680712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,6144,0.009157333109113906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,2560,0.016301333904266357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,128,0.005372444374693765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,3072,0.019204444355434842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,64,0.0017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,7168,0.014879110786649915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,16384,0.0166284441947937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,16384,0.09453777472178142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,3072,0.02626488937271966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,256,0.010055999788973067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,8192,0.024970667229758367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,32,0.001839111041691568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,768,0.011351111034552256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,1024,0.015104000767072042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,65536,0.07339377535714044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,5120,0.07982666624916925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,512,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,65536,0.12570133474138048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,12288,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,1024,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,8192,0.01073066641887029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,65536,0.06253688865237765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,64,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,65536,0.07413511143790351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,5120,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,16384,0.21795554955800375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,768,0.016181333197487723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,2560,0.004905777672926585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,8192,0.011428444749779172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,6144,0.12356000476413304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,128,0.0018702221827374564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,1024,0.005629333357016246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,4096,0.013101333545313941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,6144,0.018894222047593858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,4096,0.06798311074574788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,10240,0.036488887336519026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,128,0.002124444478087955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,5120,0.04236444499757555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,6144,0.018481777773963082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,65536,0.06547644403245714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,65536,0.17267199357350668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,512,0.002505777817633417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,2048,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,2048,0.02964622113439772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,2560,0.004543111142185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,12288,0.19232000244988334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,2048,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,256,0.02429155508677165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,3072,0.06567466921276517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,256,0.0022195554855797025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,7168,0.07721955246395536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,8192,0.07043200069003634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,512,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,65536,0.1176604429880778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,10240,0.015625778171751235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,10240,0.015818667080667283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,10240,0.02604622145493825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,7168,0.021778666310840186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,2560,0.009301332963837517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,32,0.0018177777528762817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,7168,0.017493334081437852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,3584,0.01682577696111467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,3584,0.016955556141005624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,3584,0.021935999393463135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,10240,0.8689955605400933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,512,0.015848888291252982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,12288,0.00792622235086229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,256,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,4096,0.05837955739763048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,10240,0.09649955564075047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,768,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,6144,0.04101422097947862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,1536,0.016725333200560678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,3072,0.007911110917727152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,4096,0.439846224255032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,8192,0.01976799964904785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,256,0.016504888733228047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,128,0.015389333168665567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,8192,0.05649955405129326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,6144,0.02832711074087355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,16384,0.05284444491068522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,65536,0.13477333386739096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,2560,0.008657777474986183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,256,0.0052888890107472735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,256,0.007676444119877285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,4096,0.11587733692593044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,4096,0.009016888837019602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,7168,0.07361600134107801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,12288,0.02404088940885332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,8192,0.07337422503365411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,64,0.007311111523045435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,1536,0.00865955568022198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,256,0.013798221945762634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,2048,0.006130666782458623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,2048,0.006362666686375936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,768,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,65536,0.2549617820315891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,768,0.008725333544943068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,256,0.012038222617573209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,512,0.003257777748836411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,10240,0.06556889083650377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,64,0.005442666510740916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,8192,0.057468447420332164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,512,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,3072,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,12288,0.3049200110965305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,16384,0.07422222031487359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,32,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,32,0.0021217777911159727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,768,0.0035093331502543557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,4096,0.009744000103738572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,65536,0.0788266658782959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,2560,0.007575111256705389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,3584,0.019507555498017203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,8192,0.011696000066068439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,3072,0.01218844453493754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,5120,0.008656000097592672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,2048,0.006296000132958095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,3584,0.011687111523416309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,7168,0.062077330218421094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,3584,0.008824889030721452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,12288,0.05968088573879666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,2560,0.037631111012564764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,6144,0.0411484440167745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,7168,0.03427822060055203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,8192,0.010674667027261523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,32,0.0018053334206342697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,3072,0.009294222626421187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,2048,0.01584533353646596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,1536,0.006632888896597757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,128,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,1536,0.003536000019974179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,16384,0.8785911136203342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,2048,0.009393778112199571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,6144,0.024715556038750544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,8192,0.01570311188697815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,10240,0.012387555506494312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,64,0.010135110881593492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,32,0.058876441584693066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,3584,0.04435644547144572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,10240,0.010662222074137794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,1024,0.017459554804695975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,65536,0.0738453335232205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,768,0.007162666983074612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,16384,0.025768889321221247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,2560,0.03420711225933499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,256,0.002546666603949335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,3072,0.007297777467303806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,3584,0.027440889014138117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,16384,0.026486222942670185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,5120,0.03866577810711331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,768,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,128,0.014090667168299357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,256,0.00350222239891688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,7168,0.020524443851576913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,1536,0.029976887835396662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,10240,0.03627644313706292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,1024,0.003152888889114062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,1536,0.0058755555914507965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,256,0.018230222993426852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,512,0.010259555445777046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,16384,0.033626665671666466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,3584,0.01441422270403968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,3072,0.07042488786909315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,1024,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,768,0.005631111148330901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,768,0.009230222139093611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,65536,0.09077866872151692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,8192,0.18191199832492402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,128,0.0144213338692983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,32,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,768,0.005614222337802251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,1024,0.013864888913101621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,4096,0.008339555727110969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,16384,0.03628266519970364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,32,0.01553600033124288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,32,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,512,0.021983999345037673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,3584,0.013519110778967539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,12288,0.023972445064120825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,7168,0.013233777549531726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,10240,0.012523555921183692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,7168,0.02294755478700002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,6144,0.029324445459577773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,256,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,6144,0.027595554788907368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,32,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,1024,0.00315022220214208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,3072,0.007483555210961237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,5120,0.009296000003814697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,64,0.002475555572244856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,1536,0.10322933064566718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,512,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,1024,0.012548444171746572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,4096,0.02091911103990343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,4096,0.0170817772547404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,1024,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,6144,0.02197244432237413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,1024,0.01548266741964552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,2048,0.042805333932240806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,256,0.003136000078585413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,256,0.024719999896155462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,2560,0.018570666511853535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,7168,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,128,0.012630222572220696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,8192,0.0425146652592553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,2560,0.006904888898134232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,7168,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,12288,0.21345688237084282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,1536,0.004559111264016894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,768,0.012114666402339935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,3072,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,3584,0.00423200014564726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,4096,0.007391110890441471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,10240,0.28712712393866646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,5120,0.018245332770877414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,65536,0.08214044570922852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,128,0.014747555057207743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,7168,0.01038933296998342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,1536,0.037315554089016385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,5120,0.1143875519434611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,2560,0.0069955554273393415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,3072,0.009405333134863112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,4096,0.022127999199761286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,10240,0.02772444486618042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,12288,0.012085333466529846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,3584,0.008640000389681922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,10240,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,2560,0.006688888702127669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,2048,0.031242665317323472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,2560,0.00488711107108328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,6144,0.01869511107603709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,10240,0.01478666729397244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,65536,0.17237422201368544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,1536,0.015840000576443143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,5120,0.009584888815879822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,64,0.025064888927671645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,1024,0.0057653333577844836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,3584,0.015832000308566622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,768,0.0148062225845125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,1024,0.05738400088416206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,768,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,7168,0.04815555612246195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,16384,0.025090666280852422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,7168,0.6954879760742188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,3072,0.008223999705579545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,12288,0.18882755438486734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,2048,0.9630301793416342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,1024,0.00423911131090588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,5120,0.020406223005718656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,16384,0.08593510919147068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,768,0.006934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,16384,0.1407502227359348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,3584,0.010627555350462595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,1024,0.018901333212852478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,8192,0.018968888454967074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,1536,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,64,0.002159111201763153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,5120,0.03819110989570618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,3072,0.02286311156219906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,64,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,8192,0.030231111579471167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,3584,0.01069155583779017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,8192,0.03341155582004123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,8192,0.02132266594303979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,7168,0.0052728888889153796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,10240,0.018232888645595975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,12288,0.017479999197853934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,6144,0.14135732915666369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,2048,0.015437333120240105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,8192,0.0893484420246548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,32,0.005894222193294101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,4096,0.009373333719041612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,2048,0.017667555146747164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,3072,0.1648577716615465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,32,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,2560,0.015842666228612263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,8192,0.011049778097205691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,512,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,7168,0.0414737794134352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,1536,0.01408266690042284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,512,0.004226666771703296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,128,0.0031537777847713898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,4096,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,8192,0.011402666568756104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,8192,0.021577777134047613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,256,0.0151555554734336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,3584,0.009034666750166152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,128,0.0147724449634552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,768,0.007287110719415877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,256,0.00981777740849389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,65536,0.24203732278611925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,128,0.017855111095640395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,128,0.007974222302436829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,3072,0.024494222468800012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,128,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,2560,0.01680888897842831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,12288,0.021388444635603163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,3072,0.009339555270142024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,256,0.015448888142903646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,256,0.006903999795516332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,7168,0.17518399821387398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,768,0.0120328888297081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,2048,0.01074488874938753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,7168,0.03126666612095303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,2560,0.03216800093650818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,3072,0.00832177781396442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,512,0.002554666664865282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,8192,0.015499555402331881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,32,0.0018320000833935207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,64,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,65536,0.060560888714260526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,16384,0.03816533419820998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,128,0.006182222316662471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,4096,0.020949333906173706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,2560,0.009304000271691216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,1024,0.025945777694384258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,1024,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,3584,0.011214222345087262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,16384,0.03916177815861172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,12288,1.8032080332438152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,768,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,3072,0.012732444538010491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,128,0.008612444831265343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,65536,0.05657511287265354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,6144,0.5949609014723036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,2560,0.03687377770741781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,512,0.00795911086930169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,2560,0.03386755453215705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,64,0.009988444546858469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,6144,0.029677334758970473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,512,0.002554666664865282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,12288,0.047114666965272695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,7168,0.012786666552225748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,2560,0.018168000711335074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,2560,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,3584,0.03882488939497206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,1536,0.00655644428398874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,2560,0.007458666960398356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,7168,0.005191111316283544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,128,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,512,0.0028640000770489373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,512,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,2560,0.01369599997997284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,3584,0.016576000385814242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,128,0.05050933361053467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,256,0.28508710861206055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,5120,0.03793066740036011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,1024,0.016194666425387066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,1024,0.015077332655588785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,256,0.014919999572965833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,6144,0.012410666379663678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,12288,0.15134310722351074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,6144,0.032925334241655134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,12288,0.04251377781232198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,512,0.005401777724424998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,768,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,8192,0.026178666286998328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,65536,0.06095021963119507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,128,0.011968889170222811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,10240,0.020052444603708055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,1024,0.015040889382362366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,6144,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,32,0.02680533296532101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,10240,0.045893334680133395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,2048,0.7023350927564832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,5120,0.017092444830470614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,3072,0.013036444783210754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,3072,0.020510221521059673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,1536,0.008596444295512306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,64,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,5120,0.08473422129948933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,8192,0.044396443499459155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,65536,0.17045422395070395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,512,0.016064888901180692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,2048,0.007558222446176741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,1024,0.007664889097213745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,4096,0.008676444490750631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,512,0.006575111299753189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,1536,0.010680889089902243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,5120,0.008967999782827165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,64,0.0018373332503769135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,2560,0.009399111072222391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,3072,0.059876441955566406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,16384,0.021923555268181696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,16384,0.017535999417304993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,7168,0.025964443882306416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,8192,0.028817776176664565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,1024,0.0029760001020299066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,128,0.034733331865734525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,65536,0.5388826794094509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,12288,0.024716445141368445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,128,0.015114666687117683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,8192,0.011016888750924004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,4096,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,2560,0.010008888940016428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,1536,0.0038284444146686127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,10240,0.08303111129336886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,8192,0.024274667104085285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,512,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,256,0.009722666607962714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,10240,0.0797226693895128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,32,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,4096,0.013855111267831592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,1024,0.07537955707973905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,8192,0.019262222780121695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,1536,0.016866667403115165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,8192,0.005610666755172942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,2560,0.006979555719428592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,2560,0.016944888565275405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,4096,0.046394666035970054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,16384,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,7168,0.08033688863118489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,4096,0.027436445156733196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,1536,0.012097777591811286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,3072,0.015492444237073263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,768,0.0028168888141711554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,4096,0.03256266646915012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,10240,0.034493333763546415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,768,0.006595555692911148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,2560,0.015464888678656684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,8192,0.013421333498424955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,256,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,8192,0.024323556158277724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,256,0.053802665736940175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,3584,0.007612444460391998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,8192,0.013370667066839008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,2560,0.013263111313184103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,16384,0.02349244389269087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,64,0.0075333333677715724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,8192,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,32,0.006594666590293248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,1024,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,1536,0.005038222091065513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,2048,0.011572444604502784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,768,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,128,0.01518311103185018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,65536,0.02372177773051792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,1024,0.004409777621428172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,16384,0.04598133431540596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,4096,0.008664000365469191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,128,0.015455999308162265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,65536,0.14838933944702148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,256,0.02759733299414317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,2560,0.02643644478585985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,128,0.031445334355036415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,32,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,65536,3.137930764092339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,65536,0.10832355419794719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,3584,0.016421332955360413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,256,0.0021520000365045336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,3072,0.07106400198406644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,65536,0.08681777450773452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,4096,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,128,0.00828888929552502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,12288,0.019299555155966017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,10240,0.03334844443533156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,128,0.016156444946924847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,10240,0.0162444445821974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,4096,0.02444177700413598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,3072,0.010230222509966956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,10240,0.036700444089041814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,6144,0.010442666709423065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,2560,0.007328000333574083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,2560,0.17753333515591094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,32,0.0018648888087934917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,3584,0.017052445146772597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,4096,0.004567999806669024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,512,0.30720888243781197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,3584,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,7168,0.011397333608733283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,3072,0.006976889239417181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,3584,0.03955555624432034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,1536,0.010795555180973478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,3072,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,6144,0.009668444593747457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,768,0.004034666551484002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,10240,0.016303999556435477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,256,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,768,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,7168,0.014404444230927361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,2560,0.016509332590632968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,128,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,512,0.017493334081437852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,10240,0.01679022279050615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,512,0.005274666680230035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,3584,0.015448000695970325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,12288,0.008430221842394935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,3072,0.014122666584120857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,128,0.0021297778520319196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,3072,0.03046222196684943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,5120,0.06141777833302816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,1024,0.004571555389298333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,3584,0.04683377676539951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,128,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,32,0.007918222082985772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,256,0.010770666930410596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,8192,0.033180445432662964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,3072,0.014782221780882942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,3584,0.012111999922328524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,1024,0.009328888522254096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,3072,0.016143111719025504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,4096,0.007947555846638149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,6144,0.028020444843504164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,6144,0.010027555955780877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,5120,0.017431111799346078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,128,0.01481066644191742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,2560,0.01796533332930671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,2560,0.011363555159833698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,7168,0.051989333497153394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,1536,0.006698666761318843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,768,0.003931555483076307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,10240,0.01237511138121287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,3584,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,7168,0.057499554422166615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,5120,0.04390666551060147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,65536,0.06423111094368829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,768,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,768,0.005285333428117964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,768,0.003576888806290097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,32,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,1536,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,512,0.0028311111446883944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,7168,0.02679022153218587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,10240,0.05314222309324476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,128,0.005320000151793162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,65536,0.07215199867884318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,65536,0.05435199870003594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,2560,0.007636444436179266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,512,0.014474666780895658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,3584,0.013667555318938361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,3072,0.012464000119103326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,5120,0.009695111049546136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,2048,0.015512888630231222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,1536,0.025495110286606684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,10240,0.06162133481767443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,3072,0.012657778130637275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,6144,0.009691555466916826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,2560,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,65536,0.15593955251905653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,128,0.005828444328573015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,10240,0.040418667925728693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,1024,0.0028764444092909494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,3584,0.018224888377719455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,2560,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,2048,0.006895111252864202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,256,0.014156444205178155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,128,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,64,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,12288,0.03422755665249295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,10240,0.012063110868136087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,12288,0.10760621892081366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,1024,0.5445795589023167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,1024,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,32,0.002520888836847411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,3072,0.016979555288950603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,16384,0.10685600174797906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,1024,0.016152888536453247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,4096,0.006230222268237009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,1536,0.01315822203954061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,12288,0.030155556069480047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,1536,0.007232888705200619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,2560,0.012576888832781049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,3584,0.014103111293580798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,32,0.00796622203456031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,12288,0.013330666555298699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,1536,0.010330666270520952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,5120,0.012979555461141797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,256,0.010317333042621613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,256,0.002489777695801523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,7168,0.011728000309732227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,768,0.013758222262064615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,5120,0.02093688812520769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,5120,0.013767110804716745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,3072,0.008278222547637092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,1024,0.004926222066084544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,5120,0.02091377807988061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,10240,0.04622933268547058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,65536,0.060237334834204785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,128,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,12288,0.013468444347381592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,1536,0.006586666736337874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,256,0.0052897776994440295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,6144,0.01684533390733931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,6144,0.028231110837724473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,1536,0.006311111152172089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,4096,1.5193102094862196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,16384,0.0422017772992452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,7168,0.014811555544535318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,1536,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,65536,0.23010312186347115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,5120,0.1539200014538235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,4096,0.009771555662155151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,1024,0.007278222176763747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,2560,0.006983111302057902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,2560,0.01921777758333418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,7168,0.009648000200589498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,5120,0.015824889143308003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,128,0.007428444094128079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,128,0.005346666607591841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,10240,0.02390044430891673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,2048,0.018559111489189994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,3584,0.008400888906584846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,3584,0.010752889017264048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,65536,0.04684888985421923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,6144,0.023471999499532912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,7168,0.0053386667536364655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,256,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,12288,0.017161778277821012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,3072,0.07876533269882202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,256,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,32,0.003891555385457145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,3072,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,768,0.02720888952414195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,2048,0.015136000182893542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,4096,0.022434666752815247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,32,0.0052800000541739995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,16384,0.020844444632530212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,12288,0.02201777696609497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,512,0.021870222356584337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,512,0.02473955518669552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,32,0.0018808889306253856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,5120,0.012929778132173749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,32,0.006116444451941385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,2560,0.0075733330514695905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,512,0.03031733300950792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,768,0.01960088809331258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,6144,0.010838222172525195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,5120,0.06519377893871732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,32,0.011815110842386881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,512,0.022281777527597215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,3584,0.004535111288229625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,5120,0.008303999900817871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,3072,1.1548133426242404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,10240,0.05681955814361572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,128,0.46118667390611434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,10240,0.019144889381196763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,768,0.006488889041874144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,1536,0.014396443963050842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,128,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,128,0.005290666802061929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,32,0.0188417782386144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,8192,0.016536888149049547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,2560,0.007636444436179266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,16384,0.0603368878364563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,2560,0.007292444507280986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,16384,0.15543999936845568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,12288,0.03480622172355652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,1536,0.018759111563364666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,2560,0.01832088828086853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,16384,0.03163199954562717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,7168,0.013559111290507846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,6144,0.037547555234697126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,8192,0.02090488870938619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,16384,0.021754667162895203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,512,0.0028000000036425064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,7168,0.013400889105266996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,3584,0.044808000326156616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,1024,0.11808978186713324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,10240,0.06016177601284451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,6144,0.010151999692122141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,10240,0.415720886654324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,768,0.006317333214812809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,1024,0.017602667212486267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,512,0.003212444484233856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,768,0.008041777544551426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,4096,0.024028443627887305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,7168,0.01775911119249132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,65536,0.02695288923051622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,768,0.0032231110251612137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,3584,0.01660355594423082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,6144,0.011651555697123209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,1536,0.008278222547637092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,12288,0.052095999320348106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,6144,0.619476424323188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,6144,0.03635911146799723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,3072,0.00832266691658232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,2048,0.006806222101052602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,3072,0.007295110987292395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,12288,0.01808888879087236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,768,0.004585777719815572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,768,0.0151591118839052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,1024,0.0038728887836138406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,2048,0.006268444574541516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,8192,0.023544889357354905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,1536,0.04133955637613932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,7168,0.15473333994547525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,12288,0.03556266758177016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,4096,0.013880000346236758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,7168,0.12418222427368164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,5120,0.1304195589489407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,2048,0.01167555567291048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,16384,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,3072,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,3584,0.30092888408237034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,6144,0.03198933270242479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,256,0.003145777723855443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,4096,0.011184000306659274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,7168,0.010149333212110732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,7168,0.023473777704768713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,1536,0.006625777731339137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,512,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,5120,0.012056888805495368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,4096,0.03854577740033468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,5120,0.1791235605875651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,3584,0.00793422261873881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,1536,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,2048,0.017887999614079792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,8192,0.018869333797030978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,256,0.015106666419241162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,512,0.005251555393139522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,6144,0.018962666392326355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,3072,0.01833599971400367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,4096,0.013386666774749756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,16384,0.09216977490319146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,4096,0.09405333466000027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,128,0.06193866994645861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,32,0.0019475554840432273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,5120,0.011342222491900126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,3584,0.011400889191362592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,7168,0.060362663533952504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,128,0.00656355544924736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,7168,0.0243848893377516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,512,0.01888799998495314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,3584,0.10524533192316692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,64,0.0018728888697094387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,7168,0.047968000173568726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,12288,0.01475644442770216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,768,0.017813333206706576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,8192,0.01072088877360026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,1024,0.02475999957985348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,4096,0.008654221892356873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,65536,0.08351644542482163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,128,0.003145777723855443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,16384,0.0680604444609748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,3072,0.014518222875065274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,7168,0.20618932776980928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,12288,0.029512001408471003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,3072,0.015304888288180033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,6144,0.028797331783506606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,4096,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,64,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,256,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,65536,0.14559733867645264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,3072,0.03186755710177951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,3072,0.03344710999064975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,12288,0.01344622257683012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,10240,0.012445333103338877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,2048,0.0127697777416971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,512,0.013024888932704926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,6144,0.009339555270142024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,4096,0.013150222599506378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,3072,0.01854222185081906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,3584,0.008654221892356873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,65536,0.061483555369906955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,16384,0.5325813293457031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,32,0.019133332702848647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,4096,0.011502222054534487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,2048,0.008185777399275038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,65536,0.08218577835294935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,7168,0.02125599980354309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,3072,0.00739911115831799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,768,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,4096,0.011335999601417117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,16384,0.08386311266157363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,6144,0.012538666526476542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,128,0.013065777719020844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,256,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,10240,0.2414026790195041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,32,0.01001422190003925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,1536,0.008440888590282863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,1536,0.008289777570300633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,8192,0.022094221578703985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,65536,0.06590577628877428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,1536,0.00797333319981893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,65536,0.062236441506279834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,7168,0.022294221652878657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,3584,0.09629688660303752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,10240,0.012096888489193387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,5120,0.013711111413107978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,65536,0.28186045752631295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,1536,0.06671466430028279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,2048,0.03196177879969279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,256,0.026704000102149114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,2048,0.00452977791428566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,12288,0.43938401010301376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,7168,0.05028622349103292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,65536,0.346085336473253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,16384,0.02641511129008399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,4096,0.13607199986775717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,10240,0.2661120096842448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,2048,0.007329777710967594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,3584,0.016509332590632968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,5120,0.021236444513003033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,16384,0.05926844146516588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,3072,0.008205333517657386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,12288,0.02365333338578542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,6144,0.014396443963050842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,16384,0.2988693449232313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,256,0.017515555024147034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,1536,0.006248888870080312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,10240,0.29565244250827366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,512,0.013743999931547375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,12288,0.03847644395298428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,768,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,3072,0.0188719994492001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,65536,0.019889778561062284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,12288,0.018397332893477548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,512,0.02725155486000909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,7168,0.043587555487950645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,768,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,5120,0.016158221496476066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,5120,0.029887109994888306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,3584,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,768,0.003097777772280905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,12288,0.15297510888841417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,3584,0.007620444728268518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,2048,0.14316533671485052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,65536,1.0892569224039714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,512,0.027447111076778833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,5120,0.011056000159846412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,256,0.012247111234400006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,256,0.02160088883505927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,1536,0.009504888620641496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,1024,0.0058142224119769196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,5120,0.009323555562231276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,5120,0.026465778549512226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,7168,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,512,0.008323555191357931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,7168,0.08273422055774264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,3584,0.014418666561444601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,10240,0.01512977812025282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,12288,0.0565591123369005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,65536,0.14666577180226645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,1536,0.006034666465388403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,16384,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,4096,0.03843199875619676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,7168,0.010623999767833285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,6144,0.019676445258988273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,3584,0.007660444411966536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,5120,0.016886222693655226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,3072,0.013079110946920184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,5120,0.01902755598227183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,3072,0.007977777885066139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,3072,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,2560,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,1024,0.018178666631380718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,3584,0.10932266712188721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,2560,0.007892444729804993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,128,0.005375999957323074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,256,0.004576000074545543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,2560,0.013341333303186627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,10240,0.011736888852384357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,12288,0.013439111411571503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,4096,0.004219555606444676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,512,0.00286666676402092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,3584,0.014475555883513557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,2560,0.0247688889503479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,256,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,4096,0.009381333159075843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,8192,0.012763555679056378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,16384,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,3584,0.008267555799749162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,512,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,16384,0.011508444117175208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,2560,0.0048942222363419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,1536,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,128,0.004927999857399198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,512,0.01481599940194024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,5120,0.01235466698805491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,12288,0.019800888167487252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,2048,0.006671999891599019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,4096,0.014868444866604276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,8192,0.012737777498033313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,7168,0.01962222158908844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,3072,0.023408888114823237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,16384,0.023366222778956097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,3584,1.6600942611694336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,1024,0.036329779360029436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,2560,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,512,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,1536,0.0038399998512532976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,6144,0.0121760004096561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,10240,0.009339555270142024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,6144,0.009649777577983009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,2048,0.004885333279768626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,3072,0.012077333198653327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,12288,0.017516444126764934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,256,0.004660444541109933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,512,0.005625777774386936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,65536,0.23638221952650282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,12288,0.023012444376945496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,6144,0.009409777820110321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,3072,0.00831200016869439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,3072,0.023885332875781592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,65536,0.10029155678219265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,12288,0.010178666975763109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,512,0.007255111303594377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,768,0.07104266352123685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,6144,0.011516444385051727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,256,0.022944889134830896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,5120,0.02533422244919671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,10240,0.014475555883513557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,16384,0.037258668078316584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,7168,0.01757600075668759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,65536,0.3228737778133816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,10240,0.005531555662552516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,2560,0.10523911317189534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,10240,0.2704497708214654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,5120,0.008618666893906063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,1024,0.00572622236278322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,2560,0.006583111153708563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,7168,0.02885955572128296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,8192,0.014434667097197639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,1536,0.006308444258239534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,7168,0.31165867381625706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,64,0.0021359999146726397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,10240,0.02083911167250739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,2560,0.004910222358173794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,6144,0.40671200222439235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,10240,0.07296622461742826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,2048,0.03174488743146261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,12288,0.027120888233184814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,10240,0.013797333670987023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,2560,0.026103110777007207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,768,0.005816000203291575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,64,0.005271111097600725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,2560,0.008996444443861643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,256,0.015127999915017022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,16384,0.043041778935326465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,1024,0.014998222390810648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,256,0.00286311118139161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,4096,0.014764444695578681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,128,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,2048,0.10225600004196167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,12288,0.017522666189405654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,5120,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,256,0.017159110969967313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,16384,0.016484444340070088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,3072,0.02699288891421424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,1024,0.013445333474212222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,2560,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,3584,0.007669332954618666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,3584,0.008678221868144141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,12288,0.019840000404251944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,32,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,6144,0.05355377660857307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,256,0.003175111073586676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,256,0.01458755632241567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,512,0.013384888569513956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,3584,0.02149244480662876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,256,0.04527022110091316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,3584,0.00797155582242542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,4096,0.01347733371787601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,512,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,8192,0.05929511123233371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,64,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,3584,0.00489066665371259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,128,0.011821333732869891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,12288,0.014456889695591398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,2048,0.04754844307899475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,8192,0.0827831096119351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,6144,0.012605333493815528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,3072,0.006955555743641323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,16384,0.03425688876046075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,4096,0.00868088917599784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,2048,0.11981244881947835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,5120,0.0591848889986674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,10240,0.01513155632548862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,768,0.003863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,16384,0.01756444407833947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,10240,0.04225866662131416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,10240,0.032933334509531655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,2048,0.008339555727110969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,16384,0.027427555786238775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,7168,0.020928000410397846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,32,0.005965333431959152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,768,0.01649777756796943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,512,0.01273155543539259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,256,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,512,0.00564533347884814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,1536,0.02825333343611823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,8192,0.011176000038782755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,10240,0.01604622271325853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,4096,0.02774488925933838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,65536,0.07045066356658936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,256,0.002584888910253843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,4096,0.0099982221921285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,8192,0.0181688881582684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,12288,0.014411555396185981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,16384,0.034509334299299456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,1536,0.034104890293545194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,128,0.014848000473446317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,5120,0.01977422171168857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,8192,0.02314844396379259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,2048,0.009338666995366415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,7168,0.010341333018408881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,4096,0.04812533325619168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,4096,0.040454222096337214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,6144,0.010435555544164447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,12288,0.14292889171176487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,6144,0.057068447271982826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,512,0.013759999639458127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,1536,0.0064328888224230866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,5120,0.03170311119821336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,768,0.018532445033391316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,2560,0.043120000097486705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,12288,0.030079109801186457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,8192,0.03895377781656053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,12288,0.13107911745707193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,4096,0.01870666609870063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,7168,0.01889600025282966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,2048,0.03727644350793626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,1024,0.011511111425028907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,65536,0.11640444066789414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,10240,0.01789955629242791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,12288,0.03399111164940728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,3072,0.010324444207880232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,12288,0.17219911681281197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,768,0.008562666674455007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,8192,0.015077332655588785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,2560,0.025616889198621113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,2048,0.020470221837361652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,1024,0.012741333080662621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,1024,0.0031804444475306403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,5120,0.008268444074524773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,7168,0.13069600529140896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,2048,0.02027466727627648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,7168,0.013752889302041797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,2560,0.007299555672539606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,32,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,2560,0.026141333911154006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,3072,0.015491555134455362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,16384,0.016180444094869826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,5120,0.0038782221575578055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,7168,0.02075822154680888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,1536,0.018602665927675035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,7168,0.6179164250691732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,512,0.04472177889611986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,10240,0.21024621857537162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,768,0.007469333708286285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,65536,0.06330044402016534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,256,0.015354666444990369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,12288,0.04430844386418661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,512,0.01701244380739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,1536,0.012840888566441007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,4096,0.00833155545923445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,8192,0.013593778014183044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,512,0.017229333519935608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,512,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,5120,0.017609778377744887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,65536,0.12693333625793457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,1024,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,32,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,7168,0.005554666535721884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,512,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,2560,0.08143911096784803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,10240,0.02808177802297804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,4096,0.008330666356616551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,2048,0.006514666808976068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,768,0.003195555466744635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,6144,0.01941244469748603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,512,0.008270222279760573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,3584,0.03231822119818793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,7168,0.019101333287027147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,8192,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,32,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,768,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,7168,0.01958311100800832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,256,0.014801777071423002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,8192,0.014541332920392355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,6144,0.011361777782440186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,2560,0.015487111277050443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,4096,0.010951999988820819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,3584,0.007989333735571967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,7168,0.05204711026615567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,256,0.0033475555893447665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,4096,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,768,0.033992889854643084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,768,0.0068408888247278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,2560,0.027573333846198186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,128,0.01612711118327247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,8192,0.04823644293679131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,6144,0.01815822223822276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,5120,0.009670221971140968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,32,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,2560,0.004921777794758479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,1536,0.00943022221326828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,12288,0.02029244436158074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,65536,0.14478222529093424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,256,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,3584,0.00829777783817715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,6144,0.7690098020765516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,768,0.022637334134843614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,6144,0.0181368887424469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,32,0.002179555594921112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,128,0.009359111388524374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,1536,0.007271111011505127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,8192,0.0236826671494378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,12288,0.019870221614837646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,4096,0.006016888966162999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,5120,0.011400889191362592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,1024,0.0031999999450312722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,65536,0.09160533216264512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,6144,0.09877333376142715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,7168,0.33541154861450195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,2560,0.021158221695158217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,768,0.006656000183688269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,512,0.019923556182119582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,128,0.017528888252046373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,4096,0.015477332803938123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,10240,0.028406222661336262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,256,0.005225777626037598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,128,0.02222488820552826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,16384,0.18085689014858672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,128,0.00629066675901413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,512,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,10240,0.08674489127265082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,12288,0.014104000396198697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,1536,0.003855111284388436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,1024,0.008216000265545314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,768,0.03765777746836344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,128,0.006637333167923822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,64,0.008286221987671321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,12288,0.017380444539917838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,6144,0.01072799993885888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,65536,0.07733066876729329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,1536,0.0236435549126731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,2560,0.006692444284756978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,4096,0.012304888831244575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,3072,0.01423733267519209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,4096,0.008168000313970778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,3072,0.01034577770365609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,8192,0.0691173341539171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,3072,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,12288,0.10611644718382095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,1024,0.021939555803934734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,4096,0.003846222327815162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,2048,0.02910400099224514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,2048,0.004870222260554631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,768,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,10240,0.07412978013356526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,3584,0.00795199970404307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,65536,0.02870044443342421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,64,0.02567822237809499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,10240,0.01902844508488973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,4096,0.008287111090289222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,512,0.014466666513019137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,3072,0.014356444279352823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,10240,0.06785688797632854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,2048,0.006276444428496891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,6144,0.41620177692837185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,2560,0.007268444531493717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,10240,0.0211173329088423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,2048,0.00452533322903845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,3584,0.04270222120814853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,3584,0.027087110612127516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,1536,0.0053075556125905775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,256,0.0137155552705129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,7168,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,4096,0.02126577827665541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,12288,0.029785778787400987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,16384,0.026857778429985046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,2048,0.038359999656677246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,4096,0.012053333222866058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,6144,0.01147911118136512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,512,0.006246222390068903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,512,0.0028657778683635923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,16384,0.17952356073591444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,4096,0.13799999819861516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,256,0.006636444479227066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,12288,0.023984889189402264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,3072,0.12312266561720107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,512,0.005229333208666907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,6144,0.010196444061067369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,10240,0.020406223005718656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,7168,0.03628355595800612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,1024,0.01681688924630483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,1536,0.006558222075303395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,3584,0.02336888843112522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,3072,0.02362222141689724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,5120,0.042169776227739125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,10240,0.02717777755525377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,512,0.01237511138121287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,12288,0.0729608866903517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,3072,0.01352533366945055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,3584,0.01862044466866387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,32,0.0019013333237833446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,2048,0.004909333255555895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,32,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,768,0.012072888513406118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,3072,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,6144,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,1024,0.013419555293189155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,12288,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,3072,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,12288,0.040940443674723305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,512,0.011393778026103973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,6144,0.027873777680926855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,12288,0.01739199956258138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,7168,0.03822933302985297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,32,0.00590844452381134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,12288,0.06448000007205539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,10240,0.02223199937078688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,64,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,3584,0.03031377659903632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,256,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,3072,0.019671999745898776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,4096,0.028491556644439697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,16384,0.04420355624622769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,12288,0.036305778556399874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,65536,0.05706666575537788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,10240,0.037695109844207764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,2560,0.10832711060841878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,8192,0.038431111309263445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,512,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,128,0.003540444291300244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,8192,0.016188444362746347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,64,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,1536,0.004918222212129169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,2560,0.009349333743254343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,8192,0.05619288815392388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,768,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,1024,0.0029599999801980127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,3072,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,10240,0.01297155519326528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,512,0.012675555215941535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,64,0.016775111357371014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,7168,0.07962222231758966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,3072,0.012741333080662621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,1024,0.0884560015466478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,1536,0.02609066665172577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,3072,0.007336888876226213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,2048,0.018199111024538677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,16384,0.030019554826948378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,3584,0.009119110802809397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,10240,0.05842222107781304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,512,0.00388355553150177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,1536,0.02344977855682373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,2560,0.025027554896142747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,7168,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,8192,0.05330577823850843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,768,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,6144,0.010363555616802638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,5120,0.0120666664507654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,2560,0.01833777791923947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,128,0.001964444501532449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,3584,0.02927466564708286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,10240,0.025046222739749487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,1024,0.01649599936273363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,1536,0.02313422163327535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,6144,0.014103111293580798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,4096,0.005625777774386936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,10240,0.00562755556570159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,4096,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,12288,0.023590222001075745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,5120,0.011020444333553314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,256,0.0021466666625605687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,1536,0.02643999954064687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,1024,0.0028684443483750024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,12288,0.023157333334287006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,2048,0.006646222124497096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,512,0.013844444519943662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,512,0.012168889244397482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,32,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,3584,0.023802666200531855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,10240,0.015637333194414776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,4096,0.01000088867213991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,3584,0.012832000023788877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,4096,0.03608444333076477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,1536,0.011048888994587792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,12288,0.030890666776233252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,2560,0.2791128953297933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,2048,0.004565333326657613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,65536,0.0919848879178365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,10240,0.01258044441541036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,512,0.005971555494599872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,16384,0.08134488926993476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,1536,0.04318844609790378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,12288,0.014673777752452426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,6144,0.00814133303032981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,4096,0.07066311438878377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,8192,0.030525333351559106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,4096,0.009677333136399588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,3072,0.034494221210479736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,8192,0.011399110986126794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,8192,0.14558400048149958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,10240,0.06791822115580241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,256,0.007696000238259633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,512,0.022259554929203455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,64,0.014512888259357877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,64,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,256,0.005736888696750005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,512,0.060328887568579785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,6144,0.027489778068330552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,10240,0.01569422251648373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,65536,0.08766133255428737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,2560,0.044414222240448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,3584,0.014270222849316068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,3072,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,12288,0.013677332964208392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,768,0.01625155574745602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,6144,0.040251556369993426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,7168,0.0707262224621243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,8192,0.01163200040658315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,32,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,6144,0.014025777578353882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,32,0.0021911110315057966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,65536,0.12282133102416992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,12288,0.021470222208235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,8192,0.011047110789351992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,256,0.011304000185595619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,2560,0.01482399966981676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,1024,0.004554666578769684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,4096,0.00927111092540953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,256,0.0035457776652442086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,2560,0.006994666324721442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,3584,0.011050666371981302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,768,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,12288,1.2622817357381184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,64,0.01071111112833023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,2560,0.007055111229419708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,256,0.002163555473089218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,128,0.017160000072585214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,65536,0.18500178390079072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,3584,0.011140444212489657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,64,0.0018097776919603348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,16384,0.03510133425394694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,12288,0.029915554655922785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,65536,0.03792799843682183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,5120,0.014781332678265043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,5120,0.020021332634819877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,1024,0.05423466695679558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,768,0.003154666680428717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,256,0.022506665852334764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,2048,0.01036177741156684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,128,0.004039111236731212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,32,0.004915555732117759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,5120,0.025742222865422566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,2048,0.013962666193644205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,256,0.0032888888898822996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,1536,0.01484266585773892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,1024,0.016737777325842116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,2560,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,5120,0.07989244328604804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,16384,0.053611556688944496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,3072,0.18109599749247232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,10240,0.01440622243616316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,2048,0.009986666341622671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,3072,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,3072,0.009776888622177972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,256,0.008610666626029545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,4096,0.00793333351612091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,8192,0.011751111182901593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,8192,0.22624444961547852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,5120,0.008681777450773451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,8192,0.024714666936132643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,3584,0.07121511300404866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,5120,0.016376889414257474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,6144,0.06058222055435181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,1536,0.004577777865860197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,3584,0.012409778104888068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,32,0.010808889236715106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,10240,0.02458400030930837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,1536,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,1024,0.004181333300140169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,2048,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,512,0.01683999929163191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,12288,0.018333334061834548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,1024,0.05334755447175768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,4096,0.0103511114915212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,3072,0.02433066732353634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,128,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,7168,0.03439022103945414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,512,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,3584,0.005740444279379315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,256,0.01716088917520311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,6144,0.01795111099878947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,2048,0.023557333482636347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,4096,0.014072888427310519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,65536,6.149629380967881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,12288,0.03663999835650126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,4096,0.01616533266173469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,64,0.006618666566080517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,5120,0.012038222617573209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,3072,0.010925333533022137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,32,0.005274666680230035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,12288,0.03380266825358073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,7168,0.19710844092898896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,10240,0.05307822095023262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,3584,0.004917333523432414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,768,0.006132444573773279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,3584,0.018928888771269057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,2048,0.01240888900227017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,7168,0.020488000578350492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,1536,0.006622222148709827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,7168,0.011511111425028907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,512,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,1536,0.00870666652917862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,768,0.07229689094755384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,128,0.012116444607575735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,1024,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,10240,0.046819554434882275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,8192,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,65536,0.009367999931176504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,512,0.012628444366984896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,6144,0.009115555220180089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,3072,0.014279999666743808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,16384,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,12288,0.05966666671964857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,512,0.010288000106811523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,2048,0.39055199093288845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,768,0.017853332890404593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,2048,0.02405333353413476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,4096,0.013928888572586907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,768,0.03360266817940606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,3584,0.01223288890388277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,128,0.01313511116637124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,65536,0.1175653272204929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,10240,0.020568889048364427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,7168,0.027786665492587622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,6144,0.02129688858985901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,4096,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,768,0.0025439999169773525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,7168,0.02532088922129737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,2048,0.025921778546439275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,4096,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,3072,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,2560,0.01627200014061398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,6144,0.018224888377719455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,2048,0.024604444702466328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,5120,0.018015111486117046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,7168,0.01036888857682546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,12288,0.017635555730925668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,12288,0.0610791113641527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,4096,0.01517599986659156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,12288,0.04244266615973579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,6144,0.009175111022260454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,4096,0.014518222875065274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,512,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,128,0.002809777855873108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,128,0.001863111017478837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,768,0.016200888488027785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,7168,0.036561777194341026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,2048,0.2205937835905287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,7168,0.027464000715149775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,32,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,1536,0.05984977881113688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,1024,0.005973333285914526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,8192,0.03199999862247043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,4096,0.013772444592581855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,128,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,65536,0.07766222291522555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,6144,0.005962666538026597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,128,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,5120,0.010324444207880232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,12288,0.18109333515167236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,10240,0.01906488835811615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,16384,0.02308711078431871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,5120,0.008586666650242275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,8192,0.017856889300876193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,65536,0.10270400179757012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,2048,0.02093066606256697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,8192,0.022642667094866436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,3072,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,65536,0.2466977702246772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,65536,0.38804266187879777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,12288,0.019017777509159513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,16384,0.0779928896162245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,3072,0.09399555789099799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,32,0.0018293333964215384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,4096,0.012679999901188744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,128,0.01516711049609714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,7168,0.014073777529928418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,512,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,7168,0.027646222048335608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,7168,0.02392533255947961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,6144,0.009017777939637503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,10240,0.030348443322711523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,768,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,6144,0.05011111166742113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,1024,0.01201777739657296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,3584,0.007932444413503012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,8192,0.013450666434235044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,2560,0.007253333098358578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,5120,0.004864888886610667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,768,0.014105778601434497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,2560,0.013191111385822296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,2048,0.006266666783226862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,5120,0.01587466730011834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,3584,0.03530933459599813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,3072,0.0074346669846110875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,16384,0.6107680002848307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,2560,0.006974221931563483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,2560,0.004911111046870549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,8192,0.22388267517089844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,2048,0.01351111133893331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,1024,0.003273777870668305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,3072,0.007416889071464539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,4096,0.008179555336634317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,6144,0.012849777936935425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,8192,0.33074845208062065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,16384,0.02946044339074029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,10240,0.0750453339682685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,6144,0.021613332960340712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,6144,0.08985600206587051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,1024,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,2048,0.015042665931913586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,1536,0.0195795562532213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,6144,0.0158915552828047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,1536,0.03618933426009284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,6144,0.1769111156463623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,3584,0.012823999755912356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,3584,0.007651555869314406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,7168,0.009945777555306753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,3584,0.050923556089401245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,65536,0.053653332922193736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,8192,0.020607110526826646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,2048,0.011295110815101199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,65536,0.06317955255508423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,10240,0.025420443879233465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,1536,0.0035208890007601846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,1536,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,8192,0.015975111060672335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,1024,0.008300444318188561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,1536,0.03018222252527873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,1024,0.6941235330369738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,128,0.0052728888889153796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,512,0.006904888898134232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,2560,0.011737777955002256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,10240,0.011976000335481433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,2560,0.18366489145490858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,6144,0.013194666968451606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,3072,0.08233244551552667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,768,0.014494222071435718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,512,0.0086986662613021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,5120,0.018983110785484314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,768,0.01751377847459581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,65536,1.1317040125528972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,10240,0.019897777173254225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,3584,0.027299554811583623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,3072,0.009950222240553962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,128,0.019896000623703003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,3584,0.013651555611027611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,3584,0.015789333317014907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,3584,0.007293333609898885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,5120,0.0557182232538859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,512,0.01238222254647149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,512,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,3072,0.008785777621799046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,3072,0.015092444088723926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,65536,0.06262311008241442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,12288,0.020247111717859905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,128,0.013769778112570444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,512,0.00962577760219574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,5120,0.03274577856063843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,10240,0.20431821876102021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,512,0.018927110566033255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,512,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,4096,0.017257778180970084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,12288,0.04731377628114489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,4096,0.03154933452606201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,16384,0.4296391010284424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,2560,0.014069333672523499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,2560,0.014184888866212634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,128,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,512,0.0028159999185138275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,3584,0.016513778103722464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,65536,0.05956355730692545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,16384,0.01679733395576477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,1024,0.01516355574131012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,3072,0.016145777371194627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,6144,0.019925332731670804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,5120,0.012735111018021902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,7168,0.08197955290476482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,5120,0.010943111446168689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,4096,0.010974222587214576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,1536,0.021425777011447485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,2048,0.008279110822412703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,512,0.002483555633160803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,4096,0.016134222348531086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,2560,0.05882844660017225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,8192,0.057016889254252114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,7168,0.02089244458410475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,768,0.01647822227742937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,512,0.006653333289755716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,12288,0.028088887532552082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,768,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,1536,0.010011555420027839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,65536,0.012140444583363004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,7168,0.01295733369059033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,16384,0.03549955619706048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,6144,0.03737599982155694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,5120,0.008658666577604082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,65536,0.06369422541724311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,6144,0.01923733287387424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,10240,0.03304622239536709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,1536,0.013659555878904132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,65536,0.2755608823564318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,3584,0.007675555845101674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,4096,0.023141332798533972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,3072,0.014459555347760519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,16384,0.11467466089460586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,768,0.0057502223385704895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,10240,0.12266666359371609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,2560,0.022427555587556627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,10240,0.05636799997753567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,3584,0.0069004446268081665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,16384,0.013634666800498962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,32,0.002833777831660377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,6144,0.029267556137508813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,2560,0.022654222117529974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,12288,0.04119911127620273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,768,0.011024889018800525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,512,0.004100444416205089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,128,0.007839110990365347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,128,0.02032977839310964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,3072,0.007248000138335758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,16384,0.03288355469703674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,512,0.026145777768558923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,512,0.05716977516810099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,128,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,10240,0.013148444394270578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,256,0.0021955555097924336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,256,0.01682133310370975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,6144,0.00998488896422916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,4096,0.00741511086622874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,12288,0.03088266650835673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,256,0.0021600000974204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,3584,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,3072,0.033344000577926636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,65536,0.03859377900759379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,7168,0.02273955610063341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,2560,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,32,0.002489777695801523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,768,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,3584,0.008010666403505538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,512,0.006424888968467712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,8192,0.03235910998450385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,12288,0.01843644513024224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,7168,0.014125333891974555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,512,0.0052728888889153796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,65536,0.06268888711929321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,5120,0.017245334055688646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,64,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,5120,0.008974221845467886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,10240,0.01312444441848331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,5120,0.024570667081409033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,5120,0.022295110755496558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,256,0.006625777731339137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,16384,0.026131555438041687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,768,0.009481777747472128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,2048,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,8192,0.24948265817430285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,768,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,6144,0.0099982221921285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,2560,0.006999111009968652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,5120,0.009413333402739631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,128,0.01977511081430647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,4096,0.03412177827623155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,8192,0.09656177626715766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,2048,0.0366284449895223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,768,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,3584,0.008038221961922115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,768,0.0028124445428450904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,5120,0.22434398863050672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,1536,0.011742221812407175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,4096,0.03984888725810581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,3584,0.019519999623298645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,64,0.002132444539003902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,512,0.017691555950376723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,3584,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,6144,0.008980444735950893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,512,0.0021768889079491296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,10240,0.021320889393488567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,768,0.015487111277050443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,2048,0.007610666255156199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,6144,0.05547555618815952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,4096,0.03392088744375441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,16384,0.05392711030112373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,32,0.0032195556494924757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,8192,0.014371555712487964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,1024,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,4096,0.02161866592036353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,16384,0.2792186737060547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,2048,0.011131555669837527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,4096,0.015808888607554965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,10240,0.026920000712076824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,5120,0.023656888140572443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,3072,0.007326222128338284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,2560,0.010796444283591377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,512,0.006259555617968242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,256,0.018625777628686693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,7168,0.029912889003753662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,512,0.015119999647140503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,2560,0.0070062221752272705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,512,0.01348622226052814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,512,0.004613333278232151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,128,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,128,0.0052933332820733385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,12288,0.018256889449225534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,7168,0.16949244340260824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,3584,0.014752000570297241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,2560,0.00888533310757743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,2560,0.027110222313139174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,1024,0.0031724443866146933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,10240,0.02475022276242574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,1024,0.0026702222724755607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,6144,0.0243484444088406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,768,0.02163911031352149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,128,0.005624889085690181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,4096,0.010993777877754636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,3584,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,3072,0.0038808890514903595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,128,0.005985777825117111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,256,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,256,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,32,0.007720000214046902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,7168,0.034480889638264976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,512,0.018025777406162687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,3584,0.022601778308550518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,512,0.0022231110682090125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,1536,0.0038746665749284956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,2048,0.01479822231663598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,2560,0.0070222218831380205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,1536,0.009504888620641496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,128,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,3584,0.011353777514563667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,768,0.0041582224269707995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,10240,0.016024889217482675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,32,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,12288,0.053364445765813194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,1536,0.02288355595535702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,2048,0.010660444696744284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,16384,0.024141333169407312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,65536,0.14719822671678331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,6144,0.01000355515215132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,65536,0.06722755564583673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,1536,0.009651555783218807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,2560,0.0309804446167416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,65536,0.06418222188949585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,65536,0.07992266946368747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,4096,0.012710221939616732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,5120,0.015018666783968607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,6144,1.3287644916110568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,6144,0.008992888861232335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,512,0.006640888750553131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,12288,0.27002755800882977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,3072,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,8192,0.027624888552559748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,128,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,128,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,3584,0.01868355605337355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,3072,0.03818755679660373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,16384,0.3344453440772162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,7168,0.01068800025516086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,10240,0.04295288854175144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,8192,0.014324444863531323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,8192,0.03801777627733018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,10240,0.008309333688682979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,16384,1.1908524831136067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,1024,0.03168355425198873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,10240,0.020040889581044514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,64,0.002858666703104973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,1024,0.020617778102556866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,8192,0.15715022881825766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,65536,0.11501599682701959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,12288,0.20129066043429902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,3072,0.13728355036841497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,1024,0.021261332763565913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,10240,0.15606666935814753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,4096,0.008005333443482717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,3072,0.00739911115831799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,65536,0.20287111070421007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,5120,0.014463111758232117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,1536,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,128,0.002165333264403873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,128,0.003581333491537306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,10240,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,10240,0.040602667464150324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,10240,0.014845333165592618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,128,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,128,0.005290666802061929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,3072,0.004610666798220741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,1536,0.030454221698972914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,256,0.002463999928699599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,1024,0.012989333106411828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,10240,0.016890666551060148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,8192,0.06530221965577868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,512,0.011857777833938599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,2048,0.00666133314371109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,8192,0.07146933343675402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,256,0.006738666858938005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,16384,0.021556443638271753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,1536,0.018207111292415194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,3072,0.02550044490231408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,10240,0.022015111313925848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,256,0.004218666503826777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,128,0.0029120000286234748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,3072,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,768,0.007921777665615082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,3072,0.021996445126003687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,6144,0.010023111270533668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,128,0.015127110812399121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,256,0.01614755557643043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,10240,0.015837333268589444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,128,0.020785777105225455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,16384,0.047192888127432935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,16384,0.01668711172209846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,512,0.0595760014322069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,7168,0.09490311145782471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,16384,0.033281776640150286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,3584,0.01311466677321328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,3072,0.010603555374675326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,2048,0.023123555713229712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,1536,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,256,0.0025351111673646504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,6144,0.01717422240310245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,10240,0.025916443930731878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,1024,0.010318222145239512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,12288,0.023024888502226934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,6144,0.018182223041852314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,2048,0.0069004446268081665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,128,0.0018542222678661346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,6144,0.024148444334665935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,32,0.008605333666006723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,3072,0.007669332954618666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,16384,0.31858134269714355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,8192,0.1548373301823934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,4096,0.009007111191749573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,128,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,256,0.015461333923869662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,512,0.014765333798196582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,2048,0.008794666992293464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,12288,0.03768444392416213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,8192,0.013742222554153867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,256,0.012884444660610624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,3584,0.01824355622132619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,8192,0.012970666918489667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,16384,0.031180444690916274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,12288,0.03689422210057577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,3072,0.032467554012934365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,2560,0.0148044443792767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,12288,0.023377777801619634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,6144,0.05741155809826321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,5120,0.0087013335691558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,4096,0.007628444168302748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,1536,0.009762666291660732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,12288,0.013440888788965015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,8192,0.027902222341961328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,16384,0.0671244462331136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,64,0.0018373332503769135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,128,0.0018195555441909367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,5120,0.016568889220555622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,1536,0.014476444986131458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,2560,0.07896622021993001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,8192,0.0069573331210348345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,256,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,65536,0.09549422396553887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,5120,0.008601778083377415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,768,0.01718488832314809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,16384,0.03198755449718899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,1024,0.06857333580652873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,16384,0.05806755357318454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,3072,0.030596445004145306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,128,0.004577777865860197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,7168,0.027531555957264368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,8192,0.03295288814438714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,128,0.01789155602455139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,3072,0.01767200065983666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,8192,0.0262444449795617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,16384,0.040371557076772056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,3072,0.01428711083200243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,10240,0.018162666095627677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,1536,0.009906666974226633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,256,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,65536,0.14429155985514322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,4096,0.0292524430486891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,6144,0.011067555182509951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,2048,0.06712622112698026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,768,0.02980355421702067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,5120,0.08374133374955918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,128,0.006586666736337874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,3584,0.014198222094111972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,12288,0.6173982090420193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,65536,0.14668799771202937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,768,0.0038408889538711975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,5120,0.011249778171380361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,65536,0.0832479993502299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,3584,0.007790221936172909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,10240,0.015464888678656684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,65536,0.06266399886873034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,256,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,65536,1.7532800038655598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,16384,0.027427555786238775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,12288,0.032877332634396024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,5120,0.02586311101913452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,2048,0.014808000789748298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,3072,0.02091288897726271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,1024,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,2560,0.023734221855799358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,6144,0.004913777940803104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,768,0.0058782220714622075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,64,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,65536,0.06274044513702393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,512,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,128,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,128,0.002463111033042272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,5120,0.011343999869293638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,3584,0.02940000096956889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,65536,0.13055645094977483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,8192,0.024687111377716064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,8192,0.020259555843141343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,16384,0.04965600040223864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,32,0.0021804444905784395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,10240,0.01479555500878228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,1536,0.012259555359681448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,7168,0.007923555870850882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,3072,0.02647288971477085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,256,0.002522666628162066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,65536,0.0632186664475335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,768,0.003532444437344869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,3072,0.02805333336194356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,1024,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,4096,0.014360000689824423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,10240,0.01700177788734436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,768,0.018178666631380718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,2048,0.011218667030334473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,6144,0.03502577874395583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,3584,0.023593778411547344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,2048,0.012426666915416718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,512,0.005535999933878581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,6144,0.024384000235133704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,512,0.009314667019579146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,3072,0.01647644407219357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,1024,0.0032239999208185407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,8192,0.019550222489568923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,10240,0.019931554794311523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,65536,0.0273626661962933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,3072,0.015508444772826301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,2560,0.015134221977657743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,1536,0.012601777911186218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,3072,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,8192,0.020279111133681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,16384,0.1138497723473443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,2048,0.0148026661740409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,12288,0.03089244498146905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,6144,0.00981422182586458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,1024,0.012471111284361945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,512,0.05602488915125529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,6144,0.032242665688196816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,10240,0.03281777766015794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,4096,0.011296889020336999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,65536,0.08043822315004137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,16384,0.03647644321123759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,65536,0.07484622134102716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,65536,0.08157777786254883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,12288,0.018790221876568265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,3072,0.038069334295060896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,128,0.014495111174053617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,12288,4.590061187744141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,1536,0.07196355528301664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,12288,0.01751733322938283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,2048,0.0702764458126492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,2560,0.02568088968594869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,2048,0.022129777404997084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,6144,0.016885333591037326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,65536,0.12112266487545437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,256,0.0034551111360390982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,512,0.018198221921920776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,7168,0.12485422028435601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,2048,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,512,0.007243555453088548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,768,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,6144,0.027606222364637587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,3584,0.01443555619981554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,1024,0.012089778151777057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,6144,0.0071928890215026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,1024,0.013734222286277346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,2560,0.027975112199783325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,3072,0.00815466625822915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,7168,0.01870933340655433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,1024,0.02169422143035465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,4096,0.11709777514139812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,512,0.006076444354322221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,2048,0.00488800017370118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,768,0.029713776376512315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,10240,0.06036000119315254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,512,0.006568000134494569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,7168,0.014290667242474027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,2560,0.017867555220921833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,12288,0.01443911095460256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,3584,0.08048178089989556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,8192,0.023123555713229712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,8192,0.011284444067213269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,512,0.020632000433074105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,8192,0.16560533311631945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,2560,0.022006221943431433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,4096,0.26417599784003365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,6144,0.0307608875963423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,12288,0.04868177904023064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,6144,0.02382844520939721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,8192,0.010694222317801582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,65536,0.07824711004892985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,16384,0.044255110952589244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,12288,0.02057066725360023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,8192,0.08928800291485256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,2048,0.020287111401557922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,2560,0.02653955585426754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,128,0.013410666750537025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,512,0.01590133375591702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,1024,0.020583111378881667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,3072,0.00793422261873881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,16384,0.019536000159051683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,512,0.01204711116022534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,256,0.0076453329788313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,5120,0.041999111572901406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,2560,0.01458666721979777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,16384,0.013901333014170328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,256,0.0033191111352708605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,12288,0.4960666762457953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,128,0.0062977779242727495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,32,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,16384,0.07699910799662273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,2560,0.03255911005867852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,64,0.0028737777223189673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,128,0.002834666727317704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,65536,0.04667377803060743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,7168,0.019588443968031142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,1536,0.0035502223504914176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,64,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,128,0.014105778601434497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,4096,0.10818755626678467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,8192,0.007432888779375289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,128,0.009395555489593083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,5120,0.010395555860466426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,256,0.01548266741964552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,6144,0.021969777014520433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,6144,0.008986666798591614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,3072,0.14301955699920654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,6144,0.01038577738735411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,3584,0.034048000971476235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,7168,0.039000001218583845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,5120,0.01643022232585483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,1536,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,16384,0.01646399994691213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,32,0.03356711069742838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,1536,0.04442488816049364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,1024,0.0032604444358083936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,8192,0.22739466031392416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,5120,0.02666933337847392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,2048,0.006474666711356904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,4096,0.030572444200515747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,3072,0.007920888562997183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,256,0.01752800080511305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,3072,0.03144711256027222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,2560,0.00490133340160052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,8192,0.047263112333085805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,256,0.36672266324361164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,128,0.012425777812798819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,1024,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,8192,0.015489778584904142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,5120,0.008709333009190029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,5120,0.013755555782053204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,2560,0.004941333499219683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,64,0.03872799873352051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,256,0.0052284445199701525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,768,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,7168,0.019582221905390423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,256,0.016823111308945548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,8192,0.15587822596232095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,65536,0.16212799814012316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,65536,0.3471333450741238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,128,0.0110871113008923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,128,0.025430222352345783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,2048,0.047437333398395114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,16384,0.019241778386963736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,65536,0.06878666745291816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,2048,0.012553777959611682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,128,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,5120,0.018728888697094388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,12288,0.03293777836693658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,12288,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,1536,0.008644444247086843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,6144,0.011679111255539788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,1024,0.0028702221396896574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,1536,0.015331554743978711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,256,0.0021253333737452826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,3584,0.08445244365268284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,12288,0.010602666272057427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,128,0.027129777603679236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,16384,0.012364444633324942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,4096,0.013825777504179211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,768,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,3584,0.013085333009560903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,4096,0.017806222041447956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,1024,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,16384,0.0424488882223765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,1024,0.013401777380042605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,2560,0.010210666391584609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,1024,0.0028977776981062363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,2560,0.007740444607204861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,4096,0.04759822289148966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,128,0.01719466679626041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,128,0.01313866674900055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,512,0.0024951110697454875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,128,0.015494222442309061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,65536,0.3092026710510254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,16384,0.025400888588693406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,10240,0.037625776396857366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,6144,0.048503110806147255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,32,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,512,0.012039110892348819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,2048,0.0057946667075157166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,512,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,512,0.008254222571849823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,10240,0.012711111042234631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,512,0.017828444639841717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,3584,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,256,0.0035173334181308746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,512,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,10240,0.01242844429281023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,3072,0.008765333228641087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,4096,0.04487199915779961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,16384,0.06006933583153618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,64,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,12288,0.047389331791136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,2048,0.013170666992664337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,256,0.015776000089115567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,3072,0.020062221421135795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,768,0.006261333409282897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,768,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,5120,0.02961777647336324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,768,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,65536,0.10659999979866876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,2560,0.010463111102581024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,512,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,12288,0.0206666671567493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,3072,0.004968889057636261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,12288,0.27499734030829537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,7168,0.012572444147533841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,3072,0.022659555077552795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,256,0.019932443896929424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,8192,0.059043553140428334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,1024,0.03958488835228814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,6144,0.021779555413458083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,6144,0.04309689005215963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,3584,0.05084088775846693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,1536,0.0035013332962989807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,6144,0.011402666568756104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,65536,0.23273422982957628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,3072,0.015083556373914083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,2560,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,3584,0.007974222302436829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,768,0.023223999473783705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,10240,0.02091466718249851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,128,0.005618666609128316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,256,0.013339555925793119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,12288,0.017853332890404593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,2560,0.0069573331210348345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,512,0.0402453343073527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,65536,0.14408089054955378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,5120,0.010501333408885531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,2048,0.01719377769364251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,512,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,2560,0.0070462218589252895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,1024,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,3072,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,16384,0.0124177775449223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,768,0.00831200016869439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,768,0.027132444911532935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,2560,0.11239733960893418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,8192,0.011382222175598145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,8192,0.019160888261265226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,1024,0.0035208890007601846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,2048,0.006694222076071634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,8192,0.030710223648283217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,256,0.014843554960356818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,4096,0.015862221519152325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,7168,0.015017777681350708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,3072,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,768,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,64,0.008099555969238281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,7168,0.022143110632896423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,8192,0.02940000096956889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,4096,0.03492888808250427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,2048,0.012349333200189801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,16384,0.697120878431532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,12288,0.024018666810459558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,2048,0.01681866745154063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,4096,0.007763555480374231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,512,0.0052791109515561005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,64,0.004899555610285865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,6144,0.010024000373151567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,512,0.011058666639857821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,3584,0.06865955723656549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,3584,0.016607999801635742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,12288,0.01807822287082672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,10240,0.14232088459862605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,1024,0.01272266689274046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,6144,0.010288000106811523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,7168,0.013529777526855469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,5120,0.010968888799349466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,5120,0.009104000197516548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,4096,0.030124445756276447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,1536,0.13617689079708523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,32,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,8192,0.013405332962671915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,2048,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,1536,0.07264800204171075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,512,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,5120,0.008977777428097194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,7168,0.019923556182119582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,3584,0.016178665889634024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,64,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,1536,0.016713778177897137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,12288,0.3831653330061171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,8192,0.004870222260554631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,1536,0.0035146665241983202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,6144,0.10031021965874566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,10240,0.15712443987528482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,8192,0.015823110938072205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,2048,0.01852799952030182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,6144,0.014113777213626437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,1024,0.007644444704055786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,8192,0.011686222420798408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,32,0.0018364443547195857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,6144,0.009314667019579146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,64,0.02897777822282579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,2048,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,2048,0.009657777845859528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,65536,0.18008711602952746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,2048,0.003973333372010125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,8192,0.4478444523281521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,512,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,8192,0.015119110544522604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,16384,0.11385866006215413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,2048,0.15716089142693415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,1024,0.006963555183675554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,10240,0.21356533633338082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,2048,0.030051555898454454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,2048,0.04651200109057956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,256,0.011704000333944956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,32,0.002458666761716207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,1024,0.009304888546466827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,768,0.005872888697518243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,128,0.002469333302643564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,2560,0.012728888955381183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,7168,0.012723555167516073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,256,0.014470222923490735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,256,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,10240,0.004916444420814514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,6144,0.011658666862381829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,7168,0.013112000293201871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,32,0.006213333457708359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,1024,0.007308444215191736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,7168,0.05203555689917671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,65536,1.9195679558648004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,64,0.007601777712504069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,16384,0.02942399846182929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,10240,0.019504000743230183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,64,0.0021831111775504216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,12288,0.015830222103330824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,512,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,16384,0.04866933491494921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,4096,0.09610044293933445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,768,0.018567111757066514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,4096,0.8786915673149956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,64,0.0070186663005087115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,3584,0.027102222045262653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,8192,0.04132711225085788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,768,0.004193777839342753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,2048,0.007269333634111617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,12288,4.651201883951823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,16384,6.4128019544813375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,8192,0.010823999842007955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,512,0.016169778174824186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,5120,0.016294222739007738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,65536,4.797644297281901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,16384,0.025585777229732935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,512,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,4096,0.00853777759604984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,65536,0.1288479963938395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,512,0.0031804444475306403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,6144,0.01787644459141625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,16384,0.6298666530185276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,6144,0.031699554787741765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,10240,0.08746666378445095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,5120,0.05265066689915127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,6144,0.1219751172595554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,2560,0.008409777449236976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,4096,0.018037334084510803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,3584,0.009296889106432596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,128,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,2048,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,1024,0.007648888561460707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,4096,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,1536,0.004217777815130022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,2560,0.22398310237460664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,1024,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,65536,0.06469688812891643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,5120,0.0372977786593967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,5120,0.008550222549173567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,4096,0.007967111137178209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,2048,0.006868444383144379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,6144,0.009726222190592024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,65536,0.06894844770431519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,12288,0.18841866652170816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,10240,0.02117066747612423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,256,0.012356444365448423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,6144,0.03295733200179206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,3584,0.022943110929595098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,10240,0.10904266436894734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,65536,0.06334133280648126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,2048,0.014710222681363424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,7168,0.010710222025712332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,4096,0.015656888484954834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,6144,0.04921333326233757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,3072,0.00815644446346495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,8192,0.15896356105804443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,256,0.017186666528383892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,512,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,2560,0.030846221579445735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,16384,0.042993777328067355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,5120,0.008661333057615492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,2560,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,10240,0.01717511150572035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,128,0.04016444418165419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,10240,0.015579554769727917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,512,0.0025048889219760895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,5120,0.007910221815109253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,12288,0.41307110256618923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,512,0.0027866665687825945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,8192,0.032251556714375816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,2048,0.004535111288229625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,1024,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,4096,0.00853066643079122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,2048,0.0069982219073507525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,64,0.00217333332531982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,5120,0.00870222184393141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,2048,0.014683556225564746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,3072,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,5120,0.01684977776474423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,128,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,3072,0.009954666925801171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,128,0.015145777000321282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,7168,0.01587377819750044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,5120,0.0244151105483373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,65536,0.14121688736809626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,768,0.019239111079110038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,3584,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,3072,0.021048888564109802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,1024,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,6144,0.16772177484300402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,65536,0.07431999842325847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,3584,0.014109333356221518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,64,0.0018444444156355327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,65536,0.8704257541232638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,768,0.014158222410413953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,7168,0.018207111292415194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,3584,0.03219022353490194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,1536,0.010712888505723743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,12288,0.024373332659403484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,3584,0.007619555625650618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,768,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,2560,0.00835111074977451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,256,0.0035253332720862497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,256,0.005623999983072281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,8192,0.08785777621799046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,2048,0.044413334793514676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,1536,0.015188443991872998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,1024,0.026888888743188646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,10240,0.008647999829716152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,8192,0.01069866700304879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,65536,0.10464977555804783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,1024,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,1536,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,12288,0.01735822194152408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,768,0.008671999805503422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,3072,0.009029332962301042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,65536,0.0639902220831977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,4096,0.003945777813593547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,2048,0.006624889042642381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,3584,0.016215999921162922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,3584,0.029327998558680218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,5120,0.034733331865734525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,512,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,128,0.015476443701320224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,768,0.005927111125654644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,8192,0.010682666467295753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,64,0.00552977787123786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,256,0.014818666709793938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,7168,0.009862222605281407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,10240,0.013794666363133324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,256,0.03612533211708069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,256,0.03330577744377984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,7168,0.12974933783213297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,512,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,10240,0.01939644416173299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,3584,0.03376266691419814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,3072,0.0076248885856734375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,6144,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,64,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,8192,0.04858577913708157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,65536,0.06499733527501424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,256,0.005962666538026597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,3584,0.030065778228971694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,2560,0.026865777042176988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,2560,0.013179555535316467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,2560,0.011927111281288994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,8192,0.046979556481043495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,8192,0.01310755560795466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,128,0.022255111071798537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,5120,0.008264888491895463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,512,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,7168,0.013045333325862885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,128,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,2048,0.006904888898134232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,1536,0.0041751112374994485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,16384,0.019864888654814828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,1536,0.0035173334181308746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,3072,0.023911999331580266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,7168,0.012061333490742577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,768,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,512,0.00590844452381134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,6144,0.017449777987268236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,512,0.0028080000645584534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,2048,0.0365226666132609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,768,0.01590844492117564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,256,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,1536,0.016015110744370353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,12288,0.10585866371790568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,3072,0.0052897776994440295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,4096,0.035967111587524414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,5120,0.026826666461096868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,768,0.020672889219390023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,65536,0.1180151171154446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,5120,0.01571466690964169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,2560,0.0070168889231152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,16384,0.040568000740475126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,4096,0.00795555528667238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,2048,0.0123857781291008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,768,0.011151110960377587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,128,0.005208888815508948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,128,0.006384000182151794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,7168,0.03359377715322707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,3584,0.008973333570692275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,1536,0.006933333145247565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,2048,0.025663110944959853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,7168,0.014696000350846184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,65536,3.4606683519151478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,3584,0.009966221948464712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,4096,0.025047111842367385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,7168,0.052074667480256825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,16384,0.021095999413066443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,2048,0.00629333323902554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,65536,0.06713866525226168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,4096,0.007975111405054728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,512,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,3072,0.014885332849290637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,2048,0.009311999711725447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,12288,0.07165333297517565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,16384,0.021367111139827307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,16384,0.03807466559939914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,12288,0.048650665415657886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,32,0.17933244175381133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,256,0.015444444285498725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,64,0.008401778009202745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,16384,0.09249333540598552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,12288,0.18717511494954428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,2048,0.005236444373925527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,4096,0.008660444782839881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,8192,0.015504888362354703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,64,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,256,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,16384,0.30771644910176593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,12288,0.5875031153361002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,7168,0.01038044442733129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,7168,0.011049778097205691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,3584,0.011024889018800525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,32,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,16384,0.05749422311782837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,7168,0.013081777426931592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,1024,0.006255110932721033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,128,0.005249777601824866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,256,0.015116444892353482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,2560,0.020260444945759244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,128,0.005719999886221356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,7168,0.014112000664075216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,3072,0.007968000239796108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,128,0.0127688886390792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,5120,0.033357332150141396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,16384,0.1613937748803033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,32,0.003531555748648114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,12288,0.05436888999409146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,512,0.0075431110130416016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,1024,0.004495999879307217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,10240,0.04272088739607069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,256,0.014488000008794995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,768,0.015423999892340766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,2048,0.025136889682875738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,512,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,2560,0.01071200023094813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,7168,0.013408888545301227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,8192,0.01421688828203413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,128,0.005758222192525864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,768,0.0027964444210131965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,5120,0.09641066524717544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,7168,0.013784888717863293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,6144,0.023660444551044043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,4096,0.10578488641315037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,8192,0.014461333552996317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,7168,0.03387644555833604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,768,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,65536,0.0477786660194397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,4096,0.03015111221207513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,12288,0.01850844422976176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,512,0.0022026666750510535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,512,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,10240,0.00452622233165635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,4096,0.014464888307783337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,12288,0.020099555452664692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,10240,0.020575111111005146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,5120,0.0406915545463562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,5120,0.009619555539555019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,5120,0.010423999693658618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,2048,0.026693332526418898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,6144,0.01132266637351778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,768,0.003543111185232798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,7168,0.010973333484596677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,1024,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,128,0.0018622221218215094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,7168,0.08797333637873332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,128,0.005695111221737332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,2560,0.05432888865470886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,3072,0.009893333746327294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,10240,0.06372444497214423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,6144,0.009885333478450775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,10240,0.014656888114081489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,16384,0.16492266125149196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,3072,0.017679111825095285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,64,0.022089777721299067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,64,0.0018026667336622875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,65536,0.07083821958965726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,512,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,256,0.03259377678235372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,65536,0.15236266454060873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,12288,0.27728978792826336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,10240,0.02674755619631873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,1536,0.006605333338181178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,12288,0.013404444687896304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,2048,0.005689777847793367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,5120,0.013415999710559845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,7168,0.5797244707743326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,32,0.004908444566859139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,512,0.01614488826857673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,1536,0.010692444112565784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,7168,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,6144,0.01628800067636702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,8192,0.01129422254032559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,768,0.015465777781274585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,10240,0.03510133425394694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,65536,0.06140622165468004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,1536,0.0083155557513237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,3072,0.00758755538198683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,8192,0.014150222142537435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,1024,0.04761155446370443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,64,0.00868800034125646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,4096,0.007683555285135905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,12288,0.019573332534896005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,6144,0.011939555406570435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,4096,0.029659556017981634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,12288,0.12629332807328966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,3584,0.007733333441946242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,256,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,128,0.017133333616786532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,256,0.005115555392371284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,2560,0.007000889215204451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,5120,0.009320889082219865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,512,0.014454222387737699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,3072,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,1024,0.03578844335344102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,16384,0.022065778573354084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,1536,0.014284445179833306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,3072,0.04765422145525614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,2560,0.007104888558387756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,32,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,16384,0.12373510996500652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,10240,0.012162666353914471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,1024,0.00702133360836241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,65536,0.012532444463835822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,2560,0.03576622075504727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,16384,0.01308444473478529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,1536,0.7285013198852539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,1024,0.0063493334584765965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,16384,0.0738453335232205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,12288,0.01926933394538032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,2048,0.024381332927280005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,16384,0.020994666549894545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,64,0.002064888882968161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,1024,0.006597333484225803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,4096,0.02186044388347202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,8192,0.03428533342149522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,3072,0.007975999679830339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,10240,0.032280001375410296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,2048,0.005033777819739448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,16384,0.0248355550898446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,3072,0.01625155574745602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,3584,0.007676444119877285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,512,0.0058755555914507965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,512,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,1536,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,65536,0.19634932941860625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,2560,0.08207733101314969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,4096,0.014480888843536377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,768,0.018153778380817838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,128,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,3072,0.03201333350605435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,6144,0.028809779220157202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,8192,0.018583110637134977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,1024,0.003206222214632564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,65536,0.06932089063856336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,16384,0.020271110865804885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,256,0.005299555758635203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,12288,0.023265777362717524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,65536,0.0935164425108168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,768,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,1536,0.010279999838935005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,4096,0.0075902218619982404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,12288,0.5710782474941677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,16384,0.11915822823842366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,128,0.0032186667538351486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,2048,0.00981866651111179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,4096,0.03905333413018121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,5120,0.047242667939927846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,12288,0.032455109887652926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,3072,0.03198400139808655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,16384,0.05217866765128242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,1536,0.0035457776652442086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,7168,0.02291733357641432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,12288,0.004928888960017098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,3584,0.013354666531085968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,4096,0.0042053332759274375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,128,0.010871111518806882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,10240,0.015674667225943673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,12288,0.025662221842341956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,16384,0.07265155845218234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,65536,0.05137866735458374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,3072,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,12288,0.04232977827390035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,2048,0.006604444649484422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,1024,0.0038844446341196695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,4096,0.008570666942331526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,2560,0.04678933487998115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,5120,0.0921493305100335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,16384,0.046127110719680786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,12288,0.045138667027155556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,2560,0.007690666450394525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,3584,0.02718222141265869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,65536,0.06735289096832275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,16384,0.11871643861134847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,6144,0.02565333412753211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,10240,0.012728888955381183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,12288,0.017647999856207106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,3584,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,7168,0.01036977767944336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,256,0.020592000749376085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,16384,0.02027466727627648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,512,0.013757333159446716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,512,0.005960888746711943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,4096,0.029326223664813574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,4096,0.0105253333846728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,5120,0.017239999439981248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,8192,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,2048,0.016128000285890367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,2048,0.006533333410819371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,8192,0.01069155583779017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,256,0.00508266646001074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,10240,0.028262221150928076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,7168,0.011400889191362592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,3584,0.010664000279373594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,768,0.008952889177534316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,3072,0.011736888852384357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,128,0.005039111193683413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,6144,0.01365066650840971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,65536,1.6493529213799372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,1024,0.01036266651418474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,128,0.016466667254765827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,1024,0.006170666466156642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,7168,0.02952622373898824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,1536,0.024845333562956914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,1024,0.016582222448454965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,1024,0.005967999911970562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,3072,0.03351555599106683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,5120,0.01605422298113505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,1024,0.0075911109646161394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,2048,0.004606222112973531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,3072,0.007272889216740926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,6144,0.060495111677381724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,5120,0.11597332689497207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,7168,0.009723555710580613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,7168,0.08684888813230728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,7168,0.020541333489947848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,65536,0.18032088544633654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,7168,0.018608000543382432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,5120,0.008314666648705801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,1536,0.014111111561457316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,7168,0.023271110322740342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,7168,0.05760266383488973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,512,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,768,0.06252621942096286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,4096,0.014075555735164218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,7168,0.0709199971622891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,7168,0.03601510988341437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,256,0.02143200072977278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,3584,0.01687111126052009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,3584,0.03599822190072801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,2048,0.020288000504175823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,16384,0.02479644450876448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,7168,0.07424266470803155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,1024,0.00314666661951277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,6144,0.009674666656388177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,5120,0.011417778001891242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,1536,0.004874666531880696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,2560,0.016186666157510545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,7168,0.07876000139448378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,6144,0.024130667249361675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,128,0.014512000812424554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,256,0.016470222009552848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,512,0.0028355556229750314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,1536,0.009345778160625035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,128,0.020450666546821594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,5120,0.21133866575029162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,512,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,12288,0.04671377936999003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,3072,0.015793777174419828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,1024,0.015544000599119397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,3584,0.00795911086930169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,128,0.016191111670600045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,10240,0.39585065841674805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,256,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,128,0.00869422240389718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,3584,0.10820177528593276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,256,0.021295999487241108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,3584,0.029884444342719182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,2560,0.02513955533504486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,3584,0.012807110945383707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,768,0.019239111079110038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,8192,0.07082933187484741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,3584,0.02544533378548092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,12288,0.058952887852986656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,128,0.05149955550829569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,12288,0.041101333167817854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,6144,0.010038221875826517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,12288,0.01885511146651374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,3072,0.010073777702119615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,768,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,7168,0.011020444333553314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,10240,0.011316444310877057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,128,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,3584,0.020484444167878892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,64,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,3584,0.03334133492575751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,6144,0.035952889257007174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,65536,0.13146311706966823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,2560,0.014350222216712104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,10240,0.019241778386963736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,16384,1.4342471228705511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,6144,0.09992444515228271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,1536,0.011354666617181567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,12288,0.08918844328986274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,65536,0.08052266968621148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,16384,0.011363555159833698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,10240,0.027594667341974046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,1024,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,2048,0.024312888582547505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,6144,0.009660444325870937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,65536,0.06142933501137627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,3584,0.02570755614174737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,10240,0.025569778349664476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,3072,0.007327111230956183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,6144,0.013521778086821238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,64,0.010050666828950247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,7168,0.012115555504957834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,1024,0.01889333294497596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,32,0.003573333223660787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,6144,0.24437154663933647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,768,0.015974221958054435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,10240,0.08391910791397095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,65536,0.3062044514550103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,7168,0.013783999615245394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,2048,0.027450667487250432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,32,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,1536,0.006267555471923616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,768,0.015817777978049386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,4096,0.012416889270146688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,128,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,8192,0.023959999283154804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,12288,0.036687112516827054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,16384,0.038902223110198975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,65536,0.07684977849324544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,5120,0.014551111393504672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,5120,0.008608000146018134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,16384,0.055416888660854764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,5120,0.014376888672510782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,10240,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,2048,0.004541333350870344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,3072,0.024361777636739943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,1024,0.013759999639458127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,65536,0.02199199961291419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,7168,0.021189333664046392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,8192,0.031553778383466936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,3072,0.03170133299297757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,65536,0.2801022264692518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,6144,0.03371466530693902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,65536,0.07262577613194783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,768,0.02184177769554986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,3072,0.1145448949601915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,512,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,8192,0.018927110566033255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,10240,0.014765333798196582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,5120,0.017091555727852713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,1024,0.01681600014368693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,5120,0.024944000773959692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,768,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,4096,0.02643111182583703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,10240,0.05329066514968872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,3584,0.007679111427730984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,12288,0.045088890526029796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,3072,0.013891556196742587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,8192,0.016568889220555622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,2560,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,128,0.0025724443710512584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,12288,0.027457776996824477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,12288,0.035507556464936994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,65536,0.0606515540017022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,768,0.005868444426192178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,7168,0.010423999693658618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,7168,0.0317831105656094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,512,0.002467555511328909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,16384,0.1271386676364475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,5120,0.03151288959715102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,2048,0.022637334134843614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,6144,0.023319111929999456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,3072,0.013095999757448832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,768,0.017511111166742113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,10240,0.03264088763131036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,65536,0.07825599776373969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,3584,0.38997689882914227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,1536,0.02325155503220028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,128,0.016167999969588388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,4096,0.034282667769326106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,4096,0.009877333210574256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,65536,0.09021866983837551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,1024,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,32,0.006086222413513396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,1536,0.01519555515713162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,3584,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,3072,0.015804444750150044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,4096,0.020977778567208186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,7168,0.012255111502276527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,10240,0.03631733192337884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,256,0.047120889027913414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,8192,0.015064888530307345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,12288,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,4096,0.02510844502184126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,128,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,32,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,256,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,6144,0.02879022227393256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,12288,0.01718844473361969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,4096,0.020199111766285367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,65536,0.0644293361239963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,128,0.012392000191741519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,5120,0.009331555830107795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,4096,0.03178222311867608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,512,0.007439110842016008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,256,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,64,0.0018853334089120228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,10240,0.04668088754018148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,32,0.0069102222720781965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,1024,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,16384,0.05503466725349426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,6144,0.03669510947333442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,8192,0.0887697802649604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,65536,0.245469331741333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,12288,0.02182577715979682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,256,0.0021315556433465746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,2560,0.03214222192764282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,1024,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,65536,0.08739644289016724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,2048,0.052244444688161217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,128,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,6144,0.34358400768703884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,1536,0.003546666767862108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,65536,0.08811289072036743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,12288,0.019907555646366544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,128,0.006259555617968242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,65536,0.076873779296875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,3584,0.021576889687114294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,3072,0.007316444483068254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,1024,0.014090667168299357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,128,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,3072,0.009310222334331935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,128,0.023318222827381555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,1536,0.012770666844315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,1024,0.0038924444880750445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,1536,0.019665777683258057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,16384,0.024492444263564214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,8192,0.01070577816830741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,6144,0.038995554049809776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,3072,0.23881067170037162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,12288,0.19838310612572563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,768,0.017667555146747164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,7168,0.09216266870498657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,1024,0.018171555466122098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,6144,0.00757422215408749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,3584,0.007620444728268518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,6144,0.009970666633711921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,6144,0.04953866534762912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,512,0.009315555294354757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,2560,0.024806222981876794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,1536,0.012097777591811286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,512,0.4116542339324951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,5120,0.05021333363321093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,128,0.002136888810329967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,512,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,3584,0.07511377996868558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,7168,0.03653511073854234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,3072,0.1763333347108629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,12288,0.02254666719171736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,8192,0.03825689024395413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,1024,0.011886222494973077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,8192,0.017872888180944655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,8192,0.05185777942339579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,16384,0.04859288864665561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,512,0.006675555474228329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,128,0.0018524444765514797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,4096,0.00869688888390859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,2560,0.00776088900036282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,10240,0.016198222835858662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,8192,0.023542222049501207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,1536,0.005953777581453323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,3584,0.042693333493338696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,5120,0.01792533364560869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,8192,0.01564711166752709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,7168,0.019221333993805777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,768,0.0031786666562159858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,6144,0.020263110597928364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,1024,0.004818666726350784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,1024,0.026855111122131348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,12288,0.01882577770286136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,3072,0.037552889850404524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,3584,0.029112888707054987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,6144,0.018204443984561496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,65536,0.07268089056015015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,16384,0.017525333497259352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,12288,0.01615111033121745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,4096,0.009434666898515489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,16384,0.015236443943447538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,7168,0.07079644335640801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,5120,0.11892267068227132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,512,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,6144,0.02097333305411869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,1024,0.005951111101441913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,3072,0.00758755538198683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,64,0.049682666858037315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,2560,0.02294755478700002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,12288,0.021575111481878493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,4096,0.016385777129067313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,3584,0.038809776306152344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,768,0.06479644775390625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,16384,0.0244186669588089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,12288,0.02536800007025401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,8192,0.010647999743620554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,3584,0.008309333688682979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,16384,0.024009777439965144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,2560,0.006896889044178857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,3072,0.016620443926917184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,3584,0.008665777742862701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,768,0.005606222069925732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,2048,0.07411377959781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,32,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,1536,0.006193777753247156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,16384,0.032891554964913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,12288,0.039376888010236956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,1536,0.007348444726732042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,1024,0.009578666753239101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,7168,0.12352711624569362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,16384,0.026076444321208533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,5120,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,256,0.0021893332401911416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,3072,0.005943999936183293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,64,0.0032017777363459268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,10240,0.19560533099704316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,10240,0.2716195583343506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,12288,0.05787822273042467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,128,0.015136000182893542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,32,0.008978666530715095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,8192,0.02197066611713833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,256,0.00250133333934678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,65536,0.23365955882602266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,768,0.01371022231049008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,10240,0.01703644461101956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,8192,0.014785778191354541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,4096,0.023385778069496155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,256,0.006623111251327727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,768,0.007618666523032718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,768,0.008276444342401292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,256,0.002512000087234709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,4096,0.22282666630215117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,16384,0.13939021693335638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,12288,0.028794666131337483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,8192,0.008409777449236976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,6144,0.017834666702482436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,1536,0.006305777778228124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,65536,0.17467821968926323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,768,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,16384,0.03242311212751601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,768,0.005927111125654644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,3584,0.01616533266173469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,1024,0.0038702223036024305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,10240,0.023976888921525743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,3072,0.007316444483068254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,2048,0.006565333240562015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,7168,0.023445333043734234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,768,0.003863999827040566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,256,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,10240,0.03535555468665229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,65536,0.08374133374955918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,256,0.0026391111314296722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,1024,0.06283199787139893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,7168,0.019955555597941082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,2048,0.006735111276308696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,256,0.002124444478087955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,1536,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,128,0.014466666513019137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,1024,0.007301333049933116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,8192,0.03294488787651062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,128,0.07112889157401191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,128,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,32,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,2048,0.00907733374171787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,768,0.0052675555149714155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,1024,0.04586577746603224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,128,0.006298666612969504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,65536,0.38102221488952637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,2560,0.08164266745249431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,6144,0.2900826666090223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,256,0.04717955655521817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,2048,0.009003555609120263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,3072,0.025984888275464375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,2048,0.014027555783589682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,4096,0.015491555134455362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,7168,0.03154844376775954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,768,0.003044444446762403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,5120,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,5120,0.010034666293197209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,1536,0.004467555632193883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,1536,0.011870221959220039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,1536,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,8192,0.01588888963063558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,3072,0.015431111057599386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,10240,0.026888888743188646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,8192,0.016507556041081745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,256,0.0024702221983008915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,1536,0.009611555271678502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,65536,0.3159884346856011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,1536,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,6144,0.017903111047214933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,16384,0.06150666872660319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,10240,0.017198221551047433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,1024,0.006309333360857434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,768,0.008663111262851292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,12288,0.03307644526163737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,2048,0.009551111194822524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,128,0.005948444621430502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,65536,0.06589066982269287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,4096,0.6811946762932671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,1536,0.014560888210932413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,768,0.07606489128536649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,12288,0.19056977166069877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,1536,0.0321742229991489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,1024,0.014064000712500678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,512,0.01548266741964552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,1536,0.02382755610677931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,3072,0.06900622447331746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,1536,0.045773333973354764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,6144,0.025398222936524287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,65536,0.013863110707865821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,512,0.0031848889258172777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,2048,0.03465955456097921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,65536,0.13415111435784235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,768,0.017550221747822232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,65536,0.015239111251301236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,3584,0.01812000075976054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,64,0.003842666745185852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,512,0.015141333142916361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,7168,0.021593777669800654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,128,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,4096,0.03294488787651062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,1536,0.006970666348934174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,2560,0.007293333609898885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,10240,0.022856000396940444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,2048,0.00489688871635331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,5120,0.046851555506388344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,1024,0.014546665880415173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,768,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,2560,0.009336888790130615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,32,0.008615111311276754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,2560,0.016547555724779766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,512,0.01479644411140018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,7168,0.029490666257010564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,128,0.015472888946533203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,12288,0.8004453447129992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,16384,0.055663108825683594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,1024,0.009695111049546136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,12288,0.018211555149820115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,8192,0.0433573325475057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,65536,0.06346222427156237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,512,0.013092444174819522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,512,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,7168,0.03019555409749349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,16384,0.20342755317687988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,12288,0.030230220821168687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,7168,0.3956017759111192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,512,0.005428444594144821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,64,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,256,0.0058115555180443665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,12288,0.035432888401879206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,5120,0.01608444419172075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,5120,0.04440711273087395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,16384,0.052573333183924355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,2560,0.024100444383091394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,10240,0.016921778519948322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,16384,0.032079110542933144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,3584,0.04148177636994256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,64,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,1536,0.011779555843936073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,5120,0.010886222124099731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,16384,0.04787288771735298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,256,0.0060159998635451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,16384,0.08390311400095622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,128,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,1024,0.0127697777416971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,256,0.001828444500764211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,3072,0.01275199982855055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,768,0.015821332732836407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,16384,0.08430577649010552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,5120,0.01847288840346866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,65536,0.06287999947865804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,4096,0.010650667051474253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,10240,0.1060604453086853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,768,0.009011555876996782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,64,0.005209777918126848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,256,0.016550223032633465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,65536,0.011385777758227454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,5120,0.021672000487645466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,1536,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,5120,0.11316800117492676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,128,0.015139554937680563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,10240,0.0284186667866177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,65536,0.10655999845928615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,512,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,256,0.0021626665774318907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,2560,0.04507022102673849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,8192,0.013634666800498962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,5120,0.01891911029815674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,4096,0.02796088986926609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,1536,0.04563733273082309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,3072,0.028390222125583228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,6144,0.01294222225745519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,5120,0.016521778371598985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,64,0.0053288886944452924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,512,0.006310222049554189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,8192,0.09588888618681167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,3584,0.011898666620254517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,2048,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,1024,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,256,0.029178665743933782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,256,0.0018399999373488957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,768,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,6144,0.06939111153284709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,1536,0.025690666503376428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,32,0.002198222196764416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,128,0.008362666600280339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,65536,0.32209422853257924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,1024,0.003579555700222651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,6144,0.009903999666372934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,3072,0.010021333065297868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,8192,0.05464355481995476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,8192,0.012087999946541257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,768,0.006267555471923616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,2048,0.009295110901196798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,64,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,2048,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,5120,0.00905688852071762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,12288,0.0192248887485928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,12288,0.02405866649415758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,512,0.02365155518054962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,5120,0.014318222800890604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,768,0.003481777591837777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,10240,0.016484444340070088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,512,0.006134222365087933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,7168,0.012071111136012606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,512,0.0025191110455327565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,5120,0.046896000703175865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,256,0.015475556254386902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,256,0.006559999866618051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,3584,0.0635626647207472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,3072,0.007643555601437886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,2560,0.25081245104471844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,8192,0.01293955577744378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,64,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,512,0.002522666628162066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,4096,0.024318221542570326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,10240,0.2572631041208903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,5120,0.008986666798591614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,65536,0.06648711363474528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,256,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,6144,0.12952889336480036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,8192,0.01071999967098236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,65536,0.18459644582536486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,6144,0.026349332597520616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,12288,0.01310400002532535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,512,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,65536,0.01797777745458815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,12288,0.013184888495339287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,5120,0.030662222041024104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,10240,0.05581777625613742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,3584,0.011550222006109027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,768,0.004893333133724001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,512,0.005884444548024072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,10240,0.01759111053413815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,2560,0.01480888823668162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,64,0.005625777774386936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,2048,0.006488889041874144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,3072,0.007306666837798224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,2048,0.006490666833188798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,10240,0.016920000314712524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,10240,0.020936000678274367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,6144,0.04244622257020739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,512,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,5120,0.004261333495378494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,3072,0.04961244596375359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,5120,0.020979555116759405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,1024,0.0075902218619982404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,8192,0.011032889286677042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,1536,0.00666933341158761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,2048,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,6144,0.018947554959191214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,512,0.014900444282425774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,3072,0.015809777710172866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,65536,0.03641600079006619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,4096,0.0956453349855211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,3072,0.0075777777367168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,5120,0.023306666149033442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,32,0.015478221906556023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,7168,0.014998222390810648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,512,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,128,0.02164622147878011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,128,0.005879111174080107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,6144,0.009020444419648912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,12288,0.04952177736494276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,7168,0.01669066647688548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,6144,0.011014222270912595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,1536,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,6144,0.014574221438831754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,2560,0.010018666585286459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,12288,0.10743377606074016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,3584,0.013478221992651621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,2048,0.042994668086369835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,3584,0.028578665521409776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,1024,0.015805333852767944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,16384,0.030576888057920668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,4096,0.011418666276666852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,64,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,4096,0.0229751103454166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,8192,0.02231377859910329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,8192,0.026018665896521673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,6144,0.011734222372372946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,32,0.006924444602595435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,4096,0.03115644388728672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,3584,0.05596799982918633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,2560,0.011691555380821228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,2560,0.009327111144860586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,6144,0.12251288361019558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,8192,0.283095121383667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,10240,0.01297511077589459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,6144,0.007996444072988298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,512,0.008640000389681922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,8192,0.03455111053254869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,32,0.002485333424475458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,3072,0.021141333712471858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,128,0.013800889253616333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,64,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,7168,0.03266577919324239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,7168,0.02985244327121311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,5120,0.01680088871055179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,16384,0.01829155617290073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,4096,0.028424001402325098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,65536,0.3953546682993571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,12288,0.21104889445834688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,65536,0.09079111284679836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,10240,0.08515644735760158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,768,0.016970665918456185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,768,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,3072,0.007510222494602203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,1536,0.054084443383746676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,4096,0.03826755616399977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,3072,0.03930400146378411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,3584,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,3584,0.007643555601437886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,7168,0.02198311189810435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,6144,0.021734222769737244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,10240,0.011758222348160215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,8192,0.016199111938476562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,512,0.005526222288608551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,5120,0.01651733285850949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,4096,0.006265777680608962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,4096,0.009657777845859528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,3584,0.018460444278187223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,3072,0.029677334758970473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,2048,0.016184000505341422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,1024,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,256,0.0111191107167138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,65536,0.011531555818186866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,1024,0.01202311118443807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,6144,0.17144266764322916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,512,0.01995022263791826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,768,0.01680355601840549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,3072,0.011729777687125735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,5120,0.01367022262679206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,2048,0.041831112570232816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,768,0.00794311116139094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,2048,0.02535733415020837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,7168,0.02470933397610982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,128,0.003148444410827425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,16384,0.03161333335770501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,4096,0.0580844415558709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,10240,0.19887822204165986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,2048,0.009047110875447592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,10240,0.024114666713608637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,768,0.004204444587230682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,65536,0.05664800273047554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,2048,0.012712000144852532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,16384,0.04364533225695292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,5120,0.10882666375901963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,6144,0.00905511114332411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,2560,0.020602666669421725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,2560,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,6144,0.012083555261294046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,3584,0.011680888632933298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,6144,0.004582222137186262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,64,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,3584,0.016546666622161865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,7168,0.016590222716331482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,8192,0.02574399941497379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,256,0.007328000333574083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,16384,0.03084889054298401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,32,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,768,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,2560,0.0116462219092581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,10240,0.027979556057188246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,65536,0.08840266863505046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,512,0.012736000120639801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,768,0.016473778420024447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,2560,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,1024,0.02256800068749322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,64,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,768,0.017089777522616915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,768,0.0028560000161329904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,768,0.0032097777972618737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,7168,0.26744177606370717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,10240,0.02661777701642778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,1536,0.013396444420019785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,65536,0.6974310874938965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,6144,0.009024889104896123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,128,0.020706666840447318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,128,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,10240,0.004879111217127906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,12288,0.05071111188994514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,12288,0.012472888661755456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,768,0.35889334148830837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,2048,0.01959022217326694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,7168,0.04147911071777344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,2560,0.05236711104710897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,1536,0.009672000176376766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,5120,0.009003555609120263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,128,0.018368000785509746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,7168,0.659587542215983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,6144,0.29340267181396484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,12288,0.0337137778600057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,1536,0.006618666566080517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,12288,0.01667555504375034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,16384,0.021591999464564856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,768,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,2048,0.1551040013631185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,2560,0.004578666554556953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,2560,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,12288,0.02641777859793769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,16384,0.025229333175553217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,256,0.03875289029545254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,12288,0.02195466723706987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,8192,0.014567999376191033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,1536,0.03018488817744785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,1536,0.006640888750553131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,65536,0.2703893449571398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,16384,0.016493333710564505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,32,0.008100444244013892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,512,0.005579555614127054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,7168,0.045016888115141124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,5120,0.03608355588383145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,768,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,10240,0.01700799994998508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,8192,0.019592000378502738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,3584,0.00813600007030699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,8192,0.4129902256859674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,128,0.007894222107198503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,64,0.005626666463083691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,65536,0.1757111152013143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,512,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,1536,0.006774222271309958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,7168,0.061632891496022545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,2560,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,7168,0.023924445112546284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,10240,0.10989422268337673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,6144,0.03295911020702786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,2560,0.004924444274769889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,10240,0.1659031046761407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,4096,0.06836444139480591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,1024,0.01752711170249515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,2560,0.0321324434545305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,64,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,7168,0.06559377908706665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,2560,0.014879110786649915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,2048,0.008206222620275285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,6144,0.03768444392416213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,32,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,32,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,4096,0.03027555677625868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,2048,0.07371377944946289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,512,0.002519999941190084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,64,0.00408266650305854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,7168,0.010823999842007955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,6144,0.008991111483838823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,7168,0.009415999882751042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,128,0.014411555396185981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,1536,0.007044444481531779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,10240,0.07873600059085421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,2048,0.006745777610275481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,32,0.0021555556191338436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,4096,0.008316444853941599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,65536,0.07697688870959811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,16384,0.017322666115230985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,512,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,32,0.007629333270920648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,512,0.012698666916953193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,10240,0.01259200026591619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,32,0.001618666677839226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,65536,0.37291823493109805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,768,0.015122666954994202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,7168,0.09952355755700006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,768,0.014447111222479077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,12288,0.03977866636382209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,768,0.005326222214433882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,12288,0.020431111256281536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,512,0.00250133333934678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,6144,0.026744888888465032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,65536,0.050494223833084106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,256,0.025207999679777358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,8192,0.03790222273932563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,5120,0.004555555681387584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,8192,0.009887999958462184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,16384,0.016155555844306946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,12288,0.1610764397515191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,65536,0.025303110480308533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,2560,0.02680977847841051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,2048,0.00776622196038564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,256,0.027479999595218237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,12288,0.18764889240264893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,12288,0.03801155421468947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,16384,0.024057777391539678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,128,0.01479555500878228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,10240,0.038005332152048744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,32,0.004962666581074397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,7168,0.019183999962276883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,4096,0.017802667286660936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,3584,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,768,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,1536,0.013814222481515674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,768,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,12288,0.021617778473430212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,4096,0.028399109840393066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,2560,0.008805333740181392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,128,0.04556355542606778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,1536,0.003858666867017746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,16384,0.04459199971622891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,7168,0.4750480122036404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,2560,0.01742577718363868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,7168,0.030863109562132094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,2048,0.009111999637550777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,7168,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,3584,0.014730667074521383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,512,0.013076444466908773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,1536,0.02676800058947669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,65536,0.17647022671169707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,65536,0.0934693349732293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,1024,0.0028284444577164124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,3584,0.016200888488027785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,1536,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,3072,0.015798222687509324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,7168,0.07927378018697102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,512,0.002480888946188821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,2048,0.011075555450386472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,4096,0.06221244732538859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,3072,0.007440889047251807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,6144,0.03196000059445699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,512,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,16384,0.041760888364579946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,768,0.0024826667375034755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,512,0.0035075553589397003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,128,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,128,0.004921777794758479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,3584,0.1276808844672309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,2048,0.012608889076444836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,2048,0.009348444640636444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,1024,0.02208711041344537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,7168,0.05355199840333727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,512,0.0025333333760499954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,12288,0.3173466788397895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,2048,0.005081777771313985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,32,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,3584,0.02943377693494161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,65536,0.12589155303107366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,5120,0.017270222306251526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,256,0.008631111019187504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,5120,0.03234311276011997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,65536,0.1489528947406345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,768,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,4096,0.008286221987671321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,5120,0.027440889014138117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,8192,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,12288,0.009722666607962714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,2560,0.012376888758606382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,1536,0.007692444655630324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,2048,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,16384,0.03911022345225016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,3072,0.02237422267595927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,256,0.0028719999310043124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,7168,0.019884443945354886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,1024,0.0038355555799272326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,8192,0.0810319979985555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,7168,0.01072088877360026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,256,0.021701332595613267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,3072,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,10240,0.025557332568698462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,128,0.002481777841846148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,256,0.015806222955385845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,12288,0.009686222506894005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,6144,0.03298577666282654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,10240,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,3584,0.03594400154219733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,256,0.0025155554629034465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,1024,0.004549333204825719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,768,0.006261333409282897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,12288,0.5427778032090929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,128,0.0021955555097924336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,4096,0.007600888609886169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,8192,0.009546666509575313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,1536,0.04075555668936835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,4096,0.01309333327743742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,16384,0.012350222302807702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,5120,0.02379200028048621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,16384,0.023576888773176406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,1536,0.013265777793195514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,16384,0.046315554115507335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,3584,0.010888000329335531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,256,0.005601777798599667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,8192,0.013944889108339945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,10240,0.014888889259762235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,1024,0.018544000056054857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,32,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,768,0.003093333293994268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,3072,0.012527111503813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,64,0.005612444546487596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,6144,0.025568000144428674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,1536,0.008961777720186446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,12288,0.02509333358870612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,7168,0.054674665133158364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,3072,0.031523555517196655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,3072,0.021506667137145996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,2560,0.006711111300521427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,3584,0.008998221821255153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,256,0.02295911146534814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,8192,0.040327111879984535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,2048,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,512,0.005630222045713001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,10240,0.025391111771265667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,768,0.002814222127199173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,6144,0.24199999703301323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,2048,0.004576888763242298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,2048,0.08951289123959011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,768,0.0069848886794514125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,1024,0.006235555642180973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,8192,0.011037333144081963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,65536,0.17926755216386583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,7168,0.1601653363969591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,7168,0.020976889464590285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,2048,0.008005333443482717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,512,0.0031715554909573663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,64,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,128,0.022124444444974262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,5120,0.0588888857099745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,65536,0.07672533061769274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,2048,0.01478844384352366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,1536,0.007621333003044128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,1024,0.026426666312747534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,32,0.009014222357008193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,7168,1.382965299818251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,10240,0.029021332661310833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,10240,0.026664889521068994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,65536,0.0997040006849501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,8192,0.3524355623457167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,32,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,2560,0.01016977760526869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,2048,0.016491555505328707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,3584,0.007790221936172909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,10240,0.06110933091905382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,6144,0.02605244517326355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,2048,0.029600001043743555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,32,0.001839111041691568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,10240,0.286810663011339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,256,0.007272889216740926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,8192,0.023872888750500146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,768,0.01682844426896837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,2048,0.007141333487298753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,10240,0.024301333559883967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,1024,0.01205244412024816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,512,0.018162666095627677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,2560,0.021877333521842957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,16384,0.02276444435119629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,2048,0.08304088645511203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,1024,0.04529777831501431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,65536,0.061965333090888135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,32,0.0018240000224775737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,32,0.0018613334331247543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,65536,13.440347459581163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,8192,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,3584,0.009325332939624786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,4096,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,1024,0.015577778220176697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,65536,0.06904177533255683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,4096,0.015185778339703878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,3584,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,10240,0.05416533350944519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,5120,0.20020533932579887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,8192,0.010476444330480365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,768,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,65536,0.2822693453894721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,2560,0.006679111056857639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,768,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,7168,0.010042666561073726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,16384,0.0909786687956916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,6144,0.07130133443408541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,10240,0.0353999998834398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,16384,0.025952888859642878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,128,0.0021271111650599376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,12288,0.03741511040263706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,256,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,64,0.005208888815508948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,1536,0.019169777631759644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,7168,0.04526399903827243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,8192,0.011043555206722684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,12288,0.01312444441848331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,5120,0.008303999900817871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,12288,0.030294222964180842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,4096,0.007702222300900354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,1024,0.00563822231358952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,256,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,128,0.047100444634755455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,1536,0.08720178074306911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,512,0.01903288894229465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,7168,0.01145244472556644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,64,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,10240,0.04392533169852363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,512,0.002520888836847411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,1024,0.0058551111982928375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,128,0.011994666523403592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,128,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,512,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,8192,0.030568887790044148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,2048,0.019876443677478366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,1536,0.0288702216413286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,6144,0.01767200065983666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,3072,0.008578666382365758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,3584,0.008685333033402761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,8192,0.06229333082834879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,16384,0.012208888928095499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,768,0.053779555691613086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,7168,0.012436444560686747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,32,0.006624889042642381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,256,0.0021315556433465746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,512,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,256,0.017256889078352187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,1536,0.016360888878504436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,768,0.01959022217326694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,5120,0.024944000773959692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,6144,0.04790044493145413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,6144,0.009358222285906473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,512,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,3072,0.016499555773205228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,1536,0.003863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,5120,0.008444444172912175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,12288,0.1149324443605211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,65536,0.08249866962432861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,512,0.0047280001971456735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,768,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,16384,0.041859553919898145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,3584,0.031870222753948636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,512,0.0028524444335036804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,64,0.0019351111518012153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,4096,0.010489777558379702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,256,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,768,0.0041911109454101985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,12288,0.07668977975845337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,128,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,7168,1.5171840455796983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,10240,0.01590133375591702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,2048,0.005786666853560342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,1536,0.006647999915811751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,5120,0.009276444713274637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,16384,0.01575999955336253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,2560,0.10587555832333034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,16384,0.06989955239825778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,128,0.00637155564294921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,12288,0.0185253338681327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,6144,0.019034667147530448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,6144,0.037448889679378934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,3584,0.005207111024194294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,65536,0.11226044760810004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,64,0.0017911110901170308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,6144,0.021312000022994146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,768,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,512,0.010483555495738983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,2048,0.03938666648334927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,65536,0.08740088674757217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,7168,0.01888088881969452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,2048,0.004906666775544484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,5120,0.008665777742862701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,10240,0.020079111059506733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,1536,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,7168,0.02057422200838725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,3584,0.010031110710567897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,2048,0.014106666048367819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,6144,0.009159999589125315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,5120,0.00960622231165568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,12288,0.035876442988713585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,256,0.005961777849329843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,512,0.011565333439244164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,4096,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,65536,0.02602844436963399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,6144,0.04135466615358988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,768,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,8192,0.024541333317756653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,1024,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,512,0.0024702221983008915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,3072,0.017844445175594754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,4096,0.008581333690219456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,3584,1.252111964755588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,256,0.020075556304719713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,1024,0.02330400049686432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,5120,0.022703111171722412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,1536,0.08192888895670573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,12288,0.8254097832573785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,1024,0.02026844521363576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,3072,0.011424000064531961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,128,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,3072,0.007738666401969061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,32,0.00656088896923595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,65536,0.07642844650480482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,2048,0.012856889102194043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,6144,0.016041777200169034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,7168,0.008568888737095727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,1024,0.017532444662517972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,64,0.004535111288229625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,5120,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,10240,0.023337778117921617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,5120,0.031005332867304485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,6144,0.02065955599149068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,1024,0.006320000108745363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,8192,0.17397333516014946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,65536,0.048561778333452016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,1536,0.012755555411179861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,64,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,3072,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,256,0.004880889008442561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,1024,0.018604444132910836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,10240,0.01815200017558204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,256,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,768,0.008117333054542542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,7168,0.012411555482281579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,768,0.003908444609906939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,768,0.019511999355422128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,7168,0.013873777455753751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,5120,0.008374222450786168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,3584,0.008641777767075432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,128,0.01484266585773892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,64,0.001830222292078866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,4096,0.02273155583275689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,64,0.0018275556051068837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,7168,0.033589333295822144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,768,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,2560,0.012058667010731168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,12288,0.12578044997321233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,4096,0.02787644333309597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,512,0.0028648889727062653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,4096,0.011788444386588203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,64,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,768,0.017317333155208163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,1024,0.08533689048555161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,3584,0.007654222349325816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,512,0.005372444374693765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,32,0.009180444810125563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,512,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,2048,0.182349337471856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,4096,0.014503111441930136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,512,0.022314666046036616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,5120,0.008818666968080733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,7168,0.01442311041884952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,7168,0.01426577733622657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,2048,0.029013332393434312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,10240,0.026751110951105755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,16384,0.01649599936273363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,3584,0.007982222570313348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,10240,0.2430960072411431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,512,0.01312711089849472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,12288,0.013767110804716745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,12288,0.10480444298850165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,64,0.010399999717871347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,8192,0.4597911304897732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,5120,0.009206222163306342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,10240,0.026379555463790894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,512,0.008242666721343994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,2048,0.00795555528667238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,512,0.005953777581453323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,2048,0.015274667077594332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,16384,0.03425866696569655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,512,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,4096,0.029775109555986192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,256,0.002536000063021978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,32,0.00351200004418691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,5120,0.0052062223354975385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,12288,0.013735999663670858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,6144,0.01822933389080895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,32,0.005612444546487596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,512,0.006955555743641323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,65536,7.931795756022136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,32,0.0028293333533737394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,10240,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,512,0.016470222009552848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,4096,0.007988444632954067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,5120,0.018869333797030978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,16384,0.24960178799099395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,1024,0.014837332897716098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,768,0.008568888737095727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,6144,0.01752000053723653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,12288,0.014109333356221518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,5120,0.036152889331181846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,10240,0.01683466633160909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,2560,0.007602666815121968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,65536,0.07310133510165744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,256,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,12288,0.016927111479971144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,3072,0.037801778978771634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,1024,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,3584,0.029141333368089464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,2048,0.01311911145846049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,128,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,7168,0.009977777798970541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,768,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,2560,0.005009777843952179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,512,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,1024,0.0029395555870400537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,7168,0.011758222348160215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,256,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,3072,0.13139822747972277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,3584,0.01865244408448537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,7168,0.015136889285511441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,4096,0.051240000459882945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,7168,0.015466666883892484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,7168,0.01956266661485036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,12288,0.03187288840611776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,10240,0.024684444069862366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,768,0.009639110830095079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,8192,0.028631998433007136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,3584,0.01699555582470364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,3072,0.01512888901763492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,3072,0.007291555404663086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,1024,0.015528000063366361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,8192,0.028543111350801256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,16384,0.02366755571630266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,512,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,2048,0.007538666327794393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,5120,0.012292444705963135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,1024,0.008286221987671321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,8192,0.029346668057971533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,768,0.006993778049945831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,2560,0.024360888534122046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,3072,0.01034222212102678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,3584,0.02608888844648997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,1536,0.006578666468461354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,2560,0.08188800017038982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,4096,0.0105591110057301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,8192,0.14769244194030762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,5120,0.15455022123124865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,768,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,512,0.006235555642180973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,256,0.005221333354711533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,256,0.023578666978412207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,1024,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,8192,0.038176000118255615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,10240,0.012072888513406118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,8192,0.024960888756646052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,4096,0.008627555436558193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,2560,0.009008888569143083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,256,0.005909333212508096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,4096,0.008103110724025303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,2048,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,7168,0.013459555804729462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,10240,0.043960889180501304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,10240,0.04277333286073473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,3072,0.008583111067612966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,3584,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,1536,0.023876445160971746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,65536,0.06726488802168104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,1024,0.009716444545321995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,4096,0.23222756385803223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,6144,0.01093777765830358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,6144,0.017248888810475666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,512,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,4096,0.0296382208665212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,2048,0.00832444429397583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,12288,0.27798400984870064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,256,0.005447111195988125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,2048,0.014076444837782117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,6144,0.014412444498803882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,768,0.01648177703221639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,2048,0.01624000072479248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,2560,0.007328000333574083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,512,0.01721155477894677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,7168,0.030399110582139757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,2560,0.12386311425103082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,4096,0.036990222003724836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,1536,0.006229333165619109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,2048,0.0069413334131240845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,8192,0.033590220742755465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,12288,0.013746666411558786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,64,0.006904888898134232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,1024,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,512,0.013401777380042605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,3072,0.01647911138004727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,256,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,2560,0.010378667049937779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,8192,0.047817776600519814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,16384,0.016787555482652452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,4096,0.005277333160241445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,6144,0.07636621925565931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,3072,0.007655999726719326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,1024,0.0188737776544359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,1536,0.007302222152551015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,8192,0.025123554799291823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,3072,0.01538755496342977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,10240,0.01755466726091173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,3072,0.01035911093155543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,1536,0.0038693332009845306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,10240,0.052764445543289185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,7168,0.03427111109097799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,7168,0.009644444617960189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,2560,0.15471466382344565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,10240,0.013056000073750814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,768,0.005950222412745158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,7168,0.020675554871559143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,5120,0.023810666468408372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,12288,0.044850667317708336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,2560,0.007322666545708974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,768,0.008647110727098253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,256,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,128,0.02366488840844896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,3584,0.007958222594526079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,1536,0.014789332946141561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,10240,0.02684533264901903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,1536,0.00647555540005366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,65536,0.32293423016866046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,16384,0.022611555125978258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,12288,0.033422221740086876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,3072,0.007301333049933116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,32,0.0017866666118303935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,65536,0.015392000476519266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,16384,0.06429421901702881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,768,0.018560889694425795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,5120,0.018464000688658822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,3072,0.012585777375433179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,1024,0.03145866592725118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,768,0.008264000217119852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,7168,0.010337777435779572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,5120,0.022457778453826904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,512,0.006632888896597757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,65536,0.44150667720370823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,7168,0.06280799706776936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,128,0.005027555757098728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,1024,0.06283733579847547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,768,0.007271111011505127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,32,0.004537777768241035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,256,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,7168,0.04462577899297079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,1536,0.07531911134719849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,64,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,2048,0.02433066732353634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,3072,0.007734222544564142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,768,0.0028213332924577924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,32,0.003156444471743372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,10240,0.018928888771269057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,65536,0.06291644440756904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,3072,0.018346667289733887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,3072,0.018900444110234577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,10240,0.009672000176376766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,128,0.01033688916100396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,12288,0.02881066666709052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,5120,0.011703111231327057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,6144,0.030065778228971694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,3584,0.04382933179537455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,768,0.006224000205596288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,5120,0.024752888414594863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,10240,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,10240,0.013060444758998023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,1536,0.006993778049945831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,65536,0.17527466350131563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,256,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,1024,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,4096,0.009121778110663096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,16384,0.027648889356189307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,768,0.006625777731339137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,3584,0.04213688770929972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,12288,0.02273422148492601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,512,0.014426666829321118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,512,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,8192,0.010411555568377176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,2048,0.021151110529899597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,12288,0.676696883307563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,512,0.00792622235086229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,4096,0.10329422023561265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,3072,0.045244445403416954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,12288,0.2650551001230876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,1024,0.008199999729792276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,1536,0.01756800048881107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,5120,0.008656000097592672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,128,0.014984889162911309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,5120,0.009560888840092553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,256,0.002228444442152977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,7168,0.012044444680213928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,8192,0.03048444456524319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,32,0.0018026667336622875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,12288,0.018409777018758986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,8192,0.02149244480662876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,2048,0.005970666805903117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,4096,0.11885066827138265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,768,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,1536,0.006675555474228329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,3072,0.012710221939616732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,64,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,768,0.0027991111079851785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,3584,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,64,0.005894222193294101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,768,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,64,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,7168,0.02569155560599433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,5120,0.009295110901196798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,10240,0.14492800500657824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,6144,0.017328000730938382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,128,0.0028808888875775864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,5120,0.039473778671688504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,5120,0.010199999643696679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,65536,0.09772888819376628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,12288,0.012721777790122561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,2560,0.013412444127930535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,128,0.015790222419632804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,2048,0.013760888742076026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,128,0.005264889034960005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,512,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,256,0.014813333749771118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,16384,0.030290666553709242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,2048,0.04327555497487386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,256,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,3584,0.01312711089849472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,7168,0.021472889516088698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,10240,0.04385511080423991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,3072,0.030802667140960693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,1024,0.0196426659822464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,12288,0.004854222138722737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,5120,0.015620445211728414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,1024,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,6144,0.004930666751331753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,3584,0.01479644411140018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,512,0.01185155577129788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,128,0.012072000238630505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,3072,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,8192,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,10240,0.03439022103945414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,6144,0.011830222275522021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,3584,0.014879999889267815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,3072,0.016191111670600045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,4096,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,2560,0.09901599751578437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,65536,0.22814133432176378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,3072,0.007366221812036302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,3072,0.05320177806748284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,1024,0.007678222325113084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,256,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,256,0.005229333208666907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,2560,0.0069715554515520734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,1024,0.02653955585426754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,512,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,5120,0.017872888180944655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,512,0.005948444621430502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,2560,0.011362666885058085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,128,0.001842666624320878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,256,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,4096,0.008407110969225565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,5120,0.023767999476856653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,256,0.002516444358560774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,16384,0.035954667462242976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,3584,0.011014222270912595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,768,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,16384,0.037447998921076454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,3072,0.5270631048414443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,65536,0.07816000117195977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,128,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,8192,0.015309333801269531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,5120,0.03359199894799127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,6144,0.05942310889561971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,4096,0.02311022248533037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,65536,0.10723644495010376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,128,0.025238222546047632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,65536,0.07725510994593303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,16384,0.016519111063745286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,6144,0.03055822186999851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,1536,0.013755555782053204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,768,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,10240,0.12067466311984593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,128,0.006839111033413146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,4096,0.4449129104614258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,3584,0.005613333235184352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,7168,0.040844443771574236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,2048,0.016171556380059984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,32,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,10240,0.024711110525661047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,8192,0.036835554573271014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,512,0.0031999999450312722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,768,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,1024,0.0032106666929192017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,1024,0.009743111001120673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,6144,0.009167999857001835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,128,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,6144,0.01679999960793389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,1024,0.012058667010731168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,5120,0.004879111217127906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,1536,0.005483555710977978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,128,0.005920889063013925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,2560,0.00720000018676122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,12288,0.06251911322275798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,5120,0.02769777841038174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,8192,0.03684444559945001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,128,0.01165600038237042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,3584,0.016530666086408827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,10240,0.02569955587387085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,10240,0.012516444755925072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,2048,0.026000888811217412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,256,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,256,0.014160888062583076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,6144,0.03233511249224345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,12288,0.020081778367360432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,7168,0.13999999894036189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,128,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,8192,0.011382222175598145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,256,0.0021057778762446507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,6144,0.1566586626900567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,3072,0.015819556183285184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,2048,0.03189155459403992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,16384,0.016156444946924847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,6144,0.008551110823949179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,2560,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,512,0.01036800030204985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,1024,0.0028764444092909494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,4096,0.04050933321317037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,8192,0.014139556222491793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,64,0.0018808889306253856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,1536,0.029409776131312054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,16384,0.016857778032620747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,65536,0.08588088883294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,65536,0.044327110052108765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,16384,0.32454223102993435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,10240,0.03975911272896661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,3584,0.02026844521363576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,256,0.01716888944307963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,2560,0.006218666831652324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,4096,0.008276444342401292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,1536,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,64,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,3584,0.20979289213816324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,2560,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,256,0.008281778130266402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,3584,0.014097778333557976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,65536,0.08481599887212117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,10240,0.017343110508388944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,768,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,6144,0.013016888664828407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,2560,0.01442399952146742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,8192,0.014177777700954013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,3072,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,1536,0.011042666931947073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,2048,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,5120,0.04989511105749342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,128,0.016145777371194627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,1024,0.007639110916190677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,6144,0.1425902181201511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,65536,0.06255288918813069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,64,0.005404444618357553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,3072,0.007017778025733099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,6144,0.015457777513398064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,6144,0.169868442747328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,4096,0.1180986695819431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,2560,0.030563556485705908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,3072,0.028435554769304063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,1536,0.01276622215906779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,128,0.008651555412345463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,10240,0.03707022137112088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,256,0.002797333316670524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,1024,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,64,0.005290666802061929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,7168,0.029127111037572224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,5120,0.004877333425813251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,1536,0.010419555836253695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,12288,0.02770488957564036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,3072,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,16384,0.023399111297395494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,1024,0.015871110889646742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,256,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,1536,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,6144,0.06724266873465644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,256,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,3584,0.020647111866209242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,512,0.021298666795094807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,256,0.008278222547637092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,512,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,4096,0.013767999907334646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,2560,0.007585778004593319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,512,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,768,0.003173333282272021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,768,0.005741333381997214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,2048,0.02942755487230089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,2560,0.008656889200210571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,128,0.002946666752298673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,128,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,65536,0.08544355630874634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,256,0.032257778777016535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,2048,0.007249777515729268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,2560,0.011659555964999728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,10240,0.03499555587768555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,512,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,4096,0.014094221923086377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,1024,0.011731555892361535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,10240,0.01550044450494978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,8192,0.07157955567042033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,10240,0.09990400075912476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,4096,0.007896888587209914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,2560,0.053616000546349414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,4096,0.013669333524174161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,3072,0.03699377841419644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,768,0.009707555174827576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,512,0.005489777773618698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,3584,0.11082400215996636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,768,0.015112888481881885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,1024,0.006246222390068903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,128,0.026760000321600173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,2560,0.006976000136799282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,4096,0.018903111418088276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,7168,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,512,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,7168,0.010661332971519895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,5120,0.012727110750145383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,16384,0.05466755562358432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,10240,0.050795555114746094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,10240,0.013929777675204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,2048,0.022688888841205176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,6144,0.0255022214518653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,5120,0.012451555165979596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,12288,0.014712888333532544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,1536,0.009199110998047723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,1536,0.007335111498832703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,8192,0.02287111183007558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,8192,0.025175111161337957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,64,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,1536,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,8192,0.011404444773991903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,32,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,10240,2.203517278035482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,512,0.0351822210682763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,8192,0.010426667001512317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,5120,0.047112001313103564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,3584,0.006284444282452266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,2560,0.031472888257768415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,2048,0.01990133358372582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,65536,0.7299120161268445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,12288,0.03799644443723891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,1024,0.028853333658642236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,6144,0.010030222435792288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,128,0.02699022160636054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,5120,0.00906666699382994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,65536,0.046462221278084644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,2560,0.01613333324591319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,4096,0.007631999750932057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,10240,0.016727111405796476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,3072,0.007248889240953658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,128,0.010065777434243096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,8192,0.013753777576817406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,12288,0.013827555709415011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,8192,0.0053093334039052325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,1536,0.004927999857399198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,3584,0.030880000856187608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,6144,0.011580444044537015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,5120,0.008616000413894653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,768,0.007601777712504069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,3584,0.031650665733549334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,1024,0.01556799974706438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,256,0.0069253332912921906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,1024,0.011114666859308878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,4096,0.012461333639091916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,3584,0.027789334456125896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,8192,0.03791200121243795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,256,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,512,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,5120,0.009195555415418413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,128,0.01569422251648373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,768,0.016125332978036668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,512,0.017312889297803242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,65536,0.08714755376180013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,128,0.002487999904486868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,4096,0.016533333394262526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,2560,0.007268444531493717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,12288,0.004874666531880696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,2048,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,6144,0.06585333082411024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,3584,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,8192,0.03350222110748291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,5120,0.03764355513784621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,2048,0.02383377816942003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,65536,0.06643644306394789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,8192,0.241775115331014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,65536,0.06927644544177584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,768,0.009503111243247986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,512,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,256,0.004537777768241035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,10240,0.035070220629374184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,768,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,512,0.01221422188811832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,2560,0.016196444630622864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,16384,0.38962401284111875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,3584,0.00851288851764467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,512,0.005330666485759947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,3072,0.0075911109646161394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,16384,0.17971821626027426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,65536,0.12448710865444607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,6144,0.023669333921538457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,1024,0.003563555578390757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,2048,0.05716355641682943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,1024,0.031109333038330078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,2560,0.01753155555990007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,1536,0.007098666495747036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,3072,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,65536,0.06480622291564941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,4096,0.021939555803934734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,5120,0.017051556044154696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,32,0.005614222337802251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,5120,0.00868977771864997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,6144,0.01069155583779017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,6144,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,6144,0.031136890252431233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,2560,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,4096,0.011771555576059552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,2560,0.015615110596021017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,1536,0.009351111120647853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,5120,0.005960888746711943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,3584,0.016677333248986136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,4096,0.012065778175989786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,2048,0.015839111473825242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,32,0.008988444175985124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,1024,0.015640000502268474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,12288,0.29075378841824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,3072,0.007656888829337225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,3584,0.008632000121805403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,3584,0.01664266652531094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,2560,0.009010666774378883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,768,0.016947555873129103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,16384,0.020937777227825586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,65536,0.40472621387905544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,1536,0.010999110837777456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,2048,0.02197066611713833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,10240,0.012842666771676807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,8192,0.016943999462657504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,256,0.012720000412729053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,1536,0.008293333152929941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,1024,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,7168,0.01069155583779017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,4096,0.09541066487630208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,8192,0.05092622174157036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,256,0.0022799999763568244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,128,0.014816888504558139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,1536,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,2048,0.01459377838505639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,65536,0.23213065995110405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,8192,0.018908444378111098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,1536,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,1024,0.005724444571468566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,65536,0.1758026679356893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,16384,0.02196088929971059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,1536,0.10302755567762588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,128,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,512,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,2048,0.013918222652541267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,6144,0.01953866746690538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,3072,0.0762044456270006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,1536,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,7168,0.0809840030140347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,8192,0.028696000576019287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,512,0.014808000789748298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,16384,0.0236453331179089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,8192,0.013747555514176687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,4096,0.010132444401582083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,256,0.005234666582610872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,768,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,12288,0.030626667870415583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,3072,0.018912000788582694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,512,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,768,0.010406222608354356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,2048,0.01720355616675483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,5120,0.008264000217119852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,12288,0.013782222237851886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,128,0.013762666947311826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,7168,0.04203377829657661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,256,0.0107360002067354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,5120,0.017507554756270517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,4096,0.008014221986134848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,16384,0.06636710961659749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,3584,0.010816000401973724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,2048,0.004556444370084339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,3584,0.024625778198242188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,6144,0.009328000247478485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,3584,0.1257902251349555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,3584,0.028488000233968098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,768,0.015232000086042615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,4096,0.05763644642300076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,64,0.001847111102607515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,7168,0.019463999403847587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,3072,0.025384000606007043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,6144,0.027023111780484516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,16384,0.03353600038422479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,512,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,2048,0.021776000658671062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,128,0.0147724449634552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,3072,0.007338666253619724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,65536,0.0898924469947815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,2560,0.008719111482302347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,1024,0.003173333282272021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,10240,0.022639999787012737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,12288,0.011396444506115384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,65536,0.07302933269076876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,32,0.0018168888572189545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,3072,0.021959111094474792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,6144,0.021410667233996924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,768,0.022865777214368183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,3072,0.021964444054497614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,12288,0.06729066371917725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,3584,0.0251555558707979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,1536,0.006211555666393704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,1024,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,65536,0.04845155609978569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,2560,0.04429511229197184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,2048,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,2048,0.005006222261322869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,512,0.00656355544924736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,512,0.002522666628162066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,16384,0.055881778399149575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,512,0.002521777732504739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,16384,0.009297777381208207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,3072,0.06875289148754544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,6144,0.01406222250726488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,6144,0.017518222332000732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,3072,0.016428444120619033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,512,0.002472888885272874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,12288,0.03366933266321818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,768,0.04763288961516487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,3584,0.01108088923825158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,128,0.010060444474220276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,10240,0.023038221730126276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,10240,0.029699554045995075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,3584,0.004956444518433677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,3584,0.00959822204377916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,256,0.015132443772421943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,5120,0.010117332968446944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,5120,0.03453244434462653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,2560,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,512,0.014459555347760519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,32,0.0018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,3072,0.01370044466522005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,32,0.008976000050703684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,6144,0.004564444637960858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,10240,0.039918220705456205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,3072,0.014942222171359591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,5120,0.011760888828171624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,5120,0.044970668024486966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,5120,0.046557333734300405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,32,0.01312444441848331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,64,0.001889777680238088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,768,0.010015111002657149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,512,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,768,0.019235556324323017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,65536,0.05567555295096504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,2048,0.006614222294754452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,3072,0.029135111305448744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,7168,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,4096,0.011685333318180509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,16384,0.021324444148275588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,128,0.0053102220926019884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,8192,0.011711999773979187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,5120,0.04212888744142321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,3072,0.004871110949251386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,5120,0.021800888909233943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,10240,0.005613333235184352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,8192,0.053694221708509654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,1024,0.02568088968594869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,768,0.005664000080691443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,256,0.0063591111037466265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,1024,0.005927111125654644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,12288,0.09347644117143418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,1536,0.007703999678293864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,4096,0.006696888970004187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,3584,0.01717066599263085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,12288,0.03064889046880934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,4096,0.023775999744733173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,3072,0.009940444595283931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,2048,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,5120,0.008469333251317343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,768,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,32,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,1024,0.0069066666894488865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,512,0.013762666947311826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,128,0.02441155579355028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,10240,0.013857777747843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,5120,0.00870666652917862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,512,0.006591111007663939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,5120,0.07635200023651123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,6144,0.013105777402718862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,3584,0.008481777376598781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,16384,0.006024000131421619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,2560,0.020238222347365487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,7168,0.014612444572978549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,16384,0.05453777644369337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,128,0.004909333255555895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,128,0.005528000079923206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,1536,0.04250311189227634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,65536,0.0549813343418969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,32,0.0024666666156715816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,512,0.00314666661951277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,6144,0.010325333310498131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,3072,0.0440631111462911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,3584,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,512,0.020647111866209242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,7168,0.02517866591612498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,5120,0.012168000141779581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,12288,0.02015022271209293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,1536,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,1024,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,512,0.007035555938879649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,4096,0.11938044759962295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,5120,0.008572444319725037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,768,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,5120,0.03231911195649041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,256,0.0021466666625605687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,65536,0.05937866369883219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,4096,0.02844088938501146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,2560,0.02232444451914893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,5120,0.009610666996902889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,768,0.02427111069361369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,1024,0.017461333009931777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,3072,0.043022221989101835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,1536,0.0064204442832205035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,512,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,1536,0.019919999771647982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,16384,0.04616533385382759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,3584,0.006862222320503659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,3584,0.0195448895295461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,12288,0.10915199915568034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,32,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,512,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,16384,0.025759999950726826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,6144,0.08918222453859116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,128,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,1024,0.025599999560250178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,12288,0.06203822294871012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,128,0.006106666806671355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,7168,0.021592888567182753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,2048,0.008573333422342936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,5120,0.01277599980433782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,65536,0.01953866746690538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,3584,0.011295999917719098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,3584,0.01808533403608534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,16384,0.023031999667485554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,12288,0.028647998968760174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,7168,0.02223199937078688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,2560,0.024259555670950148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,3584,0.026471111509535048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,5120,0.64447021484375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,7168,0.030371556679407757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,4096,0.008009778128729926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,3584,0.008303999900817871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,2560,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,65536,2.58152347140842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,256,0.005416888743638992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,8192,0.022357333037588332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,3584,0.0185244447655148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,256,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,128,0.006072888771692912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,6144,0.007988444632954067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,5120,0.008953777452309927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,32,0.0021262222694026097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,4096,0.008289777570300633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,2560,0.007186666958861881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,3072,0.007594666547245449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,3072,0.010346666806273991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,256,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,256,0.01851377718978458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,4096,0.06452177630530463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,4096,0.028079999817742243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,128,0.01939022209909227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,64,0.00812000036239624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,4096,0.01090133355723487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,4096,0.009359999663299983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,7168,0.017621333400408428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,12288,0.10842844512727524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,5120,0.08301511075761583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,64,0.0038408889538711975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,6144,0.011056000159846412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,12288,0.026366222235891554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,256,0.015119999647140503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,768,0.017887999614079792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,256,0.008000888758235509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,5120,0.024480000138282776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,32,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,2560,0.04320533408059014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,1536,0.00832444429397583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,12288,0.013464888764752282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,1536,0.006565333240562015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,5120,0.008453333543406593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,65536,0.05703910854127672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,16384,0.015787555111779105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,65536,0.0933084421687656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,128,0.024977778395016987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,5120,0.016199111938476562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,8192,0.02453511125511593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,12288,0.042176889048682324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,4096,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,256,0.006296000132958095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,768,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,1024,0.13860355483161077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,1536,0.00563733321097162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,65536,0.06731644603941175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,1024,0.03320444292492337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,3072,0.017047999633683097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,10240,0.03224622209866842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,6144,0.022942221826977197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,12288,0.03850222296184964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,1536,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,7168,0.02425333360830943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,65536,0.0695813364452786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,256,0.005534222142563925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,7168,0.01753155555990007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,2560,0.016030222177505493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,512,0.005894222193294101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,65536,0.06366399923960368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,5120,0.03191377719243368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,10240,0.017282666431532968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,3072,0.026165333059098985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,3072,0.004920888692140579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,10240,0.08258489105436537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,5120,0.031657778554492526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,1536,0.004524444540341695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,2048,0.009311111436949836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,5120,0.04335911075274149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,128,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,16384,0.020709334148301017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,128,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,256,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,10240,0.013101333545313941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,12288,0.05700977643330892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,3584,0.023718222975730896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,128,0.010247999595271217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,3584,0.010614222122563256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,65536,0.06302666664123535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,6144,0.015999999311235216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,1024,0.019208888212839764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,512,0.01499466598033905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,7168,2.6519927978515625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,12288,0.13738221592373318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,16384,0.021056888831986323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,6144,0.029862221744325426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,12288,0.039641777674357094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,2048,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,768,0.017587555779351127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,64,0.1772071123123169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,256,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,6144,0.01259022206068039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,3584,0.01033866653839747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,7168,0.01959288948112064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,12288,0.06317777766121759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,12288,0.013773333695199756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,64,0.0018159999615616268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,3072,0.010660444696744284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,3584,0.01034222212102678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,2048,0.021711111068725586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,3072,0.0399982233842214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,128,0.006127111199829314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,1024,0.011336888704035016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,2560,0.012401777837011548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,6144,0.009030222064918941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,256,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,128,0.007923555870850882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,5120,0.03157155712445577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,3584,0.04569955666859945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,512,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,2048,0.02680355476008521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,2560,0.026766222384240892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,1536,0.017224889662530687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,4096,0.014147554834683737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,16384,0.056631108125050865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,4096,0.02474044428931342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,65536,0.042973332934909396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,2560,0.02068444424205356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,1024,0.021535999245113794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,3072,0.3437342113918728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,128,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,768,0.022477333744366963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,2560,0.008959111240175035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,4096,0.028951999213960435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,5120,0.027699554959932964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,65536,0.4213377899593777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,128,0.014389332797792224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,128,0.011718221836619906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,1024,0.0759520000881619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,6144,0.010997333460383944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,4096,0.02022133270899455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,65536,0.07376977470186022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,256,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,12288,0.013769778112570444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,2560,0.08807110786437988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,10240,0.016346666547987197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,256,0.04624888963169522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,7168,0.010035555395815108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,7168,0.12348355187310113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,2048,0.006690666493442323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,3584,0.00794311116139094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,6144,0.034529778692457415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,512,0.011069333387745751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,5120,0.009477333062224919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,8192,0.010687111152542962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,64,0.0018053334206342697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,256,0.002475555572244856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,512,0.007616000043021307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,4096,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,256,0.005420444326268301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,512,0.005208888815508948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,7168,0.014087110757827759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,5120,0.010592888626787396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,512,0.0199297782447603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,8192,0.031199110878838435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,512,0.05686133437686496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,16384,0.058040002981821694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,3584,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,3072,0.09058844380908543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,768,0.005668444352017508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,65536,0.14850578043195936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,4096,0.008394666843944127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,3072,0.010066666536860997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,1024,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,1024,0.02144977781507704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,4096,0.034274667501449585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,512,0.0332275562816196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,4096,0.01571111049917009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,1024,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,256,0.004560889055331548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,2560,0.005534222142563925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,1024,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,1536,0.0035502223504914176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,8192,0.1986640029483371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,64,0.0022479999396536085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,3072,0.0075128889746136135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,1024,0.003995555556482739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,10240,0.014928888943460254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,1536,0.06840178039338854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,1536,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,128,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,1536,0.0184568895234002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,3584,0.013187555803192986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,1024,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,1536,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,768,0.0032995556377702286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,8192,0.014378666877746582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,4096,0.0668462249967787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,6144,0.1732177734375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,2560,0.03831377625465393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,8192,0.08930577834447224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,2560,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,128,0.0155395550860299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,12288,0.013768889009952545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,8192,0.018552889426549275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,32,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,32,0.0053004444473319584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,1024,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,1024,0.003156444471743372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,768,0.01313066648112403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,512,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,8192,0.014195554786258273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,10240,0.02199911077817281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,512,0.01276088919904497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,7168,0.04433777928352356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,512,0.018967999352349173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,5120,0.018067555295096505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,2560,0.011417778001891242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,2048,0.011682666838169098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,8192,0.030369778474171955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,2560,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,2048,0.031488001346588135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,8192,0.024719999896155462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,65536,0.07698488897747464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,3584,0.04461155666245354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,8192,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,16384,0.03080444534619649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,32,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,256,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,32,0.005883555445406172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,7168,0.013432000246312885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,768,0.03516622384389242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,2048,0.03352177805370755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,4096,0.02380977736579047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,256,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,256,0.010935111178292168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,7168,0.017136888371573556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,7168,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,7168,0.009537777966923183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,3072,0.05062222149637011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,7168,0.0619751082526313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,256,0.053800000084771045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,3584,0.009770666559537252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,5120,0.0107360002067354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,6144,0.06721422407362196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,7168,0.050107555256949536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,256,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,2560,0.03858933183881972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,7168,0.02439555525779724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,8192,0.012764444781674279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,128,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,7168,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,16384,0.02997777859369914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,3072,0.0070222218831380205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,12288,0.020640888147883944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,7168,0.02796799937884013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,6144,0.012070222033394707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,10240,0.021426666114065383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,512,0.014137778017255994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,2560,0.03661511010593838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,512,0.009356444080670675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,5120,0.02198133369286855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,256,0.014027555783589682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,16384,0.3941715558369954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,3584,0.01720266706413693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,16384,0.012090666426552666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,512,0.004559999952713649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,8192,0.010662222074137794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,8192,0.014781332678265043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,8192,0.021892444954978094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,3584,0.01071999967098236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,32,0.005675555517276128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,3584,0.007997333175606197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,12288,0.06800711154937744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,10240,0.04458666510052151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,6144,0.009728000395827824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,8192,0.034649779399236046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,768,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,10240,0.022980444961124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,4096,0.0038337777886125776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,4096,0.00797333319981893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,128,0.002458666761716207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,4096,0.13004622194502088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,256,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,8192,0.025233777032958135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,4096,0.0074684446056683855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,128,0.02202044427394867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,8192,0.003900444342030419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,6144,0.02504355543189579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,16384,0.010122666756312052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,6144,1.0832097795274522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,65536,0.4628302256266276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,12288,0.021049777666727703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,512,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,2560,0.0069875551594628235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,32,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,128,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,32,0.0021048889805873237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,4096,0.6911679903666178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,12288,0.07030310895707872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,6144,0.021568000316619873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,8192,0.01979466610484653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,1024,0.007961778177155389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,3072,0.011708444191349877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,6144,0.009641778137948778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,1536,0.030418667528364394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,7168,0.06157777706782023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,3584,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,3584,0.004879111217127906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,6144,0.03163199954562717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,16384,0.020415999823146395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,4096,0.004506666627195147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,12288,0.0209840006298489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,12288,0.00796622203456031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,7168,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,8192,0.023003555006451074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,4096,0.00830844458606508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,8192,0.09190489186180963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,2560,0.004893333133724001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,8192,0.016522667474216886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,5120,0.10998844438129002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,3584,0.015595555305480957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,6144,0.013793778088357715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,3584,0.022691556149058874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,768,0.0028080000645584534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,7168,0.012788444757461548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,32,0.009342222577995723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,768,0.0031964443624019623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,10240,0.1130782233344184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,16384,0.08678222364849514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,6144,0.015517334143320719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,6144,0.09497511386871338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,768,0.0041164445380369825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,512,0.012085333466529846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,6144,0.016861332787407767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,2560,0.026159110996458266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,7168,0.014471999473041959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,128,0.005607111172543631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,6144,0.009704888694816166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,7168,0.013440888788965015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,12288,0.03129244512981839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,1536,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,2560,0.015769778026474845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,7168,0.013019555144839816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,16384,0.12881599532233343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,8192,0.01552888916598426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,2048,0.006870222174459034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,4096,0.018203554881943595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,3072,0.00832177781396442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,6144,0.021588444709777832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,65536,0.086444444126553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,5120,0.008665777742862701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,768,0.009385777844323052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,1536,0.008285333712895712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,16384,0.02130577796035343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,256,0.01478488908873664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,4096,0.012461333639091916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,7168,0.013651555611027611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,16384,0.07394399907853869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,8192,0.14570132891337076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,8192,0.020121778051058452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,5120,0.011722666521867117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,3072,0.09125155872768825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,1536,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,256,0.014494222071435718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,512,0.0025111111915773815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,1024,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,10240,0.2482551203833686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,1024,0.006583111153708563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,128,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,10240,0.22568710645039877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,5120,0.009280000295903947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,65536,0.05599288807974922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,1024,0.018186666899257235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,16384,0.020269332660569083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,65536,0.12500800026787653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,12288,0.04942844311396281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,5120,0.011327111058764987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,3584,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,4096,0.01443911095460256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,2560,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,6144,0.035164445638656616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,12288,0.027465777264700994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,1536,0.014461333552996317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,12288,0.03457155492570665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,7168,0.007511110769377814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,6144,0.05952622493108114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,12288,1.8152408599853516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,16384,0.039615111218558416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,1024,0.006646222124497096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,7168,0.029063112205929224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,65536,0.07069866524802314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,7168,0.030059556166330974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,12288,0.0376800000667572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,12288,0.01795022189617157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,4096,0.01000088867213991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,65536,0.059246222178141274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,10240,0.04502933224042257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,768,0.017232000827789307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,10240,0.02200444373819563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,512,0.0028728888266616394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,1024,0.00759733302725686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,7168,0.009709333380063375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,3584,0.010022222167915767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,3072,0.007629333270920648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,8192,0.022894221875402663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,6144,0.02136444383197361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,3072,0.03615377677811517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,7168,0.04157244496875339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,5120,0.013120888835854001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,8192,0.3478444417317708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,4096,0.008276444342401292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,1024,0.0069324444565508105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,1024,0.01348088930050532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,6144,0.006583111153708563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,4096,0.01684799955950843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,64,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,65536,0.07637866338094075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,3584,0.008302222523424361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,3584,0.15258399645487467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,10240,0.4611102210150824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,2560,0.02862400147649977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,512,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,128,0.0022035555707083806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,12288,0.060849779182010226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,512,0.0035146665241983202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,1024,0.006416888700591193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,32,0.0021199999998013177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,3584,0.007709333466158972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,512,0.01479822231663598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,2560,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,2048,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,7168,0.017824000782436795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,2560,0.01351466692156262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,1536,0.017245334055688646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,4096,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,4096,0.008280889027648503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,512,0.008345777789751688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,5120,0.01889333294497596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,2048,0.027710222535663184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,7168,0.027961777316199407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,4096,0.007999111380841997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,5120,0.01735377808411916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,6144,0.019023999571800232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,3072,0.013767999907334646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,512,0.012949333422713809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,32,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,768,0.015119999647140503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,8192,0.014505777094099255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,256,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,8192,0.02256533337963952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,8192,0.033569776349597506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,10240,0.01221333361334271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,10240,0.26022132237752277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,4096,0.01605688863330417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,1536,0.03127911024623447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,7168,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,1536,0.015503111812803479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,1024,0.013760888742076026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,64,0.0018115554832749898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,2560,0.4712924427456326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,1024,0.006261333409282897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,3072,0.009684444301658207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,512,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,3072,0.02793333265516493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,4096,0.009770666559537252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,4096,0.010556444525718689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,512,0.003015999992688497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,1024,0.016705777910020616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,256,0.01681155628628201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,512,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,32,0.04537689023547702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,16384,0.006549333532651265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,10240,0.02847377790345086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,512,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,768,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,1024,0.06277689006593493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,12288,0.008738666772842407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,512,0.005399999933110342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,256,0.0500204430686103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,2048,0.01716800034046173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,7168,0.011404444773991903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,1536,0.003543111185232798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,12288,0.04423111014895969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,768,0.01586933268441094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,3584,0.013817778064144982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,1536,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,3072,0.02351911034848955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,32,0.005669333454635408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,7168,0.02125866711139679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,12288,0.03028800090154012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,64,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,128,0.011031111081441244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,6144,0.009993777506881291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,768,0.0052800000541739995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,768,0.012450666891203986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,1536,0.01718311177359687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,12288,0.13193510638342962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,1536,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,1024,0.0038746665749284956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,768,0.011085333095656501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,65536,0.05063022176424662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,1024,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,2560,0.07706311013963488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,2560,0.026985777748955622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,65536,0.08338666624493069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,3584,0.009476444787449306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,128,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,2048,0.03064799971050686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,10240,0.04760800136460198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,768,0.04136711027887132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,6144,0.009324444664849175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,2560,0.011335111326641507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,4096,0.008488888541857401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,64,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,6144,0.07219644387563069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,256,0.005223111146026188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,3584,0.014826666977670459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,768,0.0028488888508743713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,4096,0.02607111136118571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,6144,0.029532445801628962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,3072,0.007628444168302748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,8192,0.015242666006088257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,7168,0.015487111277050443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,16384,0.07899644639756945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,1024,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,1024,0.009697777529557547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,2560,0.006052444378534953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,128,0.006037333359320958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,5120,0.01861422260602315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,512,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,65536,0.027862221002578735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,256,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,65536,0.09268355369567871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,3072,0.0236817780468199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,7168,0.014467555615637036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,1536,0.01686844395266639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,3072,0.00489066665371259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,1024,0.45502758026123047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,6144,0.013410666750537025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,5120,0.009458666874302758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,1536,0.018171555466122098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,65536,0.16083999474843344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,5120,0.017512889371977914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,12288,0.04408355553944906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,5120,0.05762222078111437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,6144,0.023042667243215773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,5120,0.03503911031617059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,64,0.005618666609128316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,65536,0.07065511412090726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,4096,0.03934311204486423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,2560,0.011951111257076263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,7168,0.009742221898502773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,768,0.0031848889258172777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,7168,0.010120000276300643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,7168,0.01791288952032725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,2560,0.01665155589580536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,1024,0.007809778054555257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,4096,0.07747111055586073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,10240,0.016520000166363187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,12288,0.030008888906902734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,2048,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,3072,0.016286222471131217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,6144,0.009352000223265754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,64,0.09917155239317153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,12288,0.03613777624236213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,5120,0.0583840012550354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,2560,0.004857777721352047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,6144,0.05574311150444878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,1024,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,256,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,12288,0.0052800000541739995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,4096,0.01479644411140018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,16384,0.01995733380317688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,3584,0.015384889311260648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,16384,1.511590215894911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,1536,0.0038319999972979226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,64,0.010653333531485664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,6144,0.018533334136009216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,65536,0.08400089210934109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,3072,0.018039999736679923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,65536,0.2476737764146593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,10240,0.027004443936877783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,512,0.0024764444679021835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,12288,0.10037510924869114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,768,0.008198222352398766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,256,0.006434666613737742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,16384,0.0333315564526452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,512,0.007301333049933116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,32,0.008010666403505538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,10240,0.016145777371194627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,256,0.005967111223273807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,128,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,32,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,12288,0.030113776524861652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,8192,0.025105777713987563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,2560,0.015156444576051501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,1024,0.003168888803985384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,10240,0.06349777513080172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,8192,0.02336799932850732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,3072,0.014916444818178812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,8192,0.044623110029432506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,6144,0.032237334383858576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,3584,0.011346666349305047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,5120,0.016176000237464905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,8192,0.0396399994691213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,1024,0.021570665968788996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,8192,0.011015999648306104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,512,0.0028560000161329904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,10240,0.01903022163444095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,65536,0.04832177691989475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,3584,0.035284443034066096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,256,0.005208000126812193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,1024,0.012016889121797351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,3072,0.009006222089131674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,64,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,3584,0.04562311040030586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,512,0.017522666189405654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,768,0.0024862223201327855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,256,0.016159999701711867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,10240,0.016164445214801364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,256,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,1536,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,64,0.013264888690577613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,65536,0.021088000800874498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,8192,0.013743999931547375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,1024,0.024412444896168176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,65536,0.06633333365122478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,1536,0.006649777707126405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,12288,0.03669955664210849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,12288,0.013078221844302284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,64,0.0018053334206342697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,256,0.015306666493415833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,65536,0.12608710924784342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,5120,0.01800622211562263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,3584,0.007273777491516537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,3584,0.03480888737572564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,65536,0.048941334088643394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,1024,0.015464888678656684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,128,0.008650666309727563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,512,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,3072,0.008667555948098501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,768,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,4096,0.010037333601050908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,6144,0.013212444053755866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,8192,0.04671288861168755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,3584,0.0284826656182607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,1536,0.010673777924643623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,512,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,1536,0.004743111216359668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,7168,0.06919644276301067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,6144,0.01184622198343277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,4096,0.015646222564909194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,8192,0.05443377627266777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,4096,0.005998222364319696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,32,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,6144,0.017521777086787753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,2048,0.005709333138333426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,256,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,256,0.01798488861984677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,8192,0.17782221900092232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,16384,0.06227022409439087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,1024,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,768,0.013775111072593264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,6144,0.03465066684616937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,65536,0.5049706565009223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,64,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,65536,0.017112889223628573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,5120,0.026941332552168105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,4096,0.008954666554927826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,8192,0.02437066700723436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,5120,0.011819555527634092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,5120,0.008994667066468133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,32,0.0028515555378463534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,32,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,1024,0.06261599726147123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,7168,0.010007999837398529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,512,0.01274577776590983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,32,0.0018008889423476325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,3584,0.02699822187423706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,2048,0.007608888877762689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,2560,0.010026666853162978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,256,0.003256888853179084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,5120,0.031222220924165513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,3584,0.042971554729673594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,768,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,10240,0.019007111589113872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,4096,0.022856000396940444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,2560,0.005927111125654644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,6144,0.010651555326249866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,5120,0.023652444283167522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,8192,0.010674667027261523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,2560,0.015111999379263984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,12288,0.031493332650926374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,8192,0.03073244293530782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,4096,0.014183110660976835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,6144,0.008612444831265343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,8192,0.019363555643293593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,12288,0.09310577975379096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,6144,0.031214220656288996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,3072,0.007843555675612556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,3072,0.0339937773015764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,16384,0.027263111538357202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,12288,0.029719110992219713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,6144,0.34451111157735187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,7168,0.21980978382958305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,128,0.005199999858935674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,1024,0.06747821966807048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,16384,0.021976888179779053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,4096,0.02068088948726654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,512,0.014160000615649752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,7168,0.004522666749027041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,8192,0.01958666741847992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,2048,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,16384,0.04672266708479988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,256,0.03524177935388353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,3584,0.004523555437723796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,4096,0.02539199921819899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,768,0.0028133332315418455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,10240,0.014409777190950183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,10240,0.023740443918440077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,5120,0.8872248861524793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,1024,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,1536,0.012401777837011548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,1536,0.008656000097592672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,4096,0.04712266723314921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,7168,0.01997155613369412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,7168,0.06056977642907036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,3072,0.007291555404663086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,3072,0.027224888404210407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,3584,0.02371911042266422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,7168,0.021698666943444148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,3584,0.31679733594258624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,1536,0.01794933279355367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,10240,0.00870488915178511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,1536,0.052370667457580566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,1024,0.008250666989220513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,2048,0.006666666517655055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,32,0.0017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,6144,0.029688000679016113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,2560,0.007329777710967594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,3072,0.007419555551475949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,2560,0.010304888917340172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,16384,0.030559112628300984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,3584,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,2560,0.008667555948098501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,768,0.01731911136044396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,7168,0.0403404434521993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,128,0.002132444539003902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,3584,0.007594666547245449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,8192,0.011054221954610614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,5120,0.04550933175616794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,5120,0.004552000098758274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,7168,0.040757334894604154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,768,0.018441778090265062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,2560,0.013554666605260638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,10240,0.020246222615242004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,64,0.009604444106419882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,10240,0.020813332663642038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,1536,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,4096,0.020609777834680345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,1536,0.018021333548757765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,4096,0.02054577734735277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,2560,0.010825778047243753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,10240,0.026611554953787062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,2560,0.07813688781526354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,256,0.014155555102560254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,12288,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,8192,0.03293600016170078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,3072,0.07144177622265287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,1536,0.008643555144468943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,8192,0.03327911098798116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,1536,0.007585778004593319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,3072,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,12288,0.014160000615649752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,5120,0.004917333523432414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,6144,0.012422222230169507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,64,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,3072,0.0038542221817705366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,8192,0.012861332959598966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,12288,0.007812444534566667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,32,0.005416888743638992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,2560,0.008744888835483128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,16384,0.06933422221077813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,1536,0.006638222270541721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,1024,0.014093332820468478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,1536,0.011706666813956367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,1024,0.0032293332947625052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,256,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,5120,0.008999110923873054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,3072,0.0069751110341813825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,4096,0.348452435599433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,128,0.014782221780882942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,7168,0.4878551165262858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,4096,0.027199111051029626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,256,0.004950222041871813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,8192,0.011940444509188334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,1024,0.018888889087571036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,7168,0.028184887435701158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,65536,0.06107733647028605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,64,0.0024826667375034755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,512,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,128,0.01819022165404426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,10240,0.018889778190188937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,1536,0.011939555406570435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,10240,0.011723555624485016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,6144,0.07484622134102716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,1536,0.015828443898095023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,512,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,6144,0.0107351111041175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,7168,0.010703999963071613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,2560,0.013742222554153867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,4096,0.018209778600268893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,8192,0.03706044289800856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,64,0.001967111188504431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,2048,0.018583999739752877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,4096,0.005984000033802456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,256,0.006598222172922558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,128,0.0021333334346612296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,16384,0.1525591082043118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,256,0.014473777678277759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,5120,0.008967111508051554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,64,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,768,0.018016000588734944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,8192,0.013637333280510373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,1536,0.0038417776425679526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,1024,0.007628444168302748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,32,0.004547555413511065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,4096,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,65536,0.14088177680969238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,7168,0.010363555616802638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,65536,0.17878488699595133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,1024,0.016837333639462788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,16384,0.07259733147091336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,8192,0.010673777924643623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,3072,0.025063999825053748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,3584,0.008964444200197855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,128,0.015438222222858004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,512,0.013119999733236102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,512,0.015802666544914246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,768,0.016935999194780987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,16384,0.7624382442898221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,10240,3.781721750895182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,7168,0.03218577636612786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,65536,0.2092897759543525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,2560,0.027412444353103638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,5120,0.025411556164423626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,768,0.011361777782440186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,128,0.0166631109184689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,4096,0.012519111235936483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,7168,0.01703733371363746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,32,0.002187555655837059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,128,0.002480888946188821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,12288,0.8213431040445963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,768,0.007992000215583378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,128,0.009319110876984065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,128,0.01260355528857973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,16384,0.029639111624823675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,1024,0.007621333003044128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,10240,0.029098666376537744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,8192,0.16682844691806367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,12288,0.024873778223991394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,1024,0.00775111135509279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,65536,0.13497689035203722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,65536,0.44473690456814235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,256,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,3584,0.009695111049546136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,7168,0.029086222251256306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,2048,0.01475911173555586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,768,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,3072,0.008882666627566019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,8192,0.03713422351413303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,5120,0.017758222089873422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,5120,0.008565333154466417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,2048,0.009663110805882348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,1024,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,16384,0.016951110627916124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,12288,0.10289866394466823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,12288,0.08782222535875107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,65536,0.06835110982259114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,16384,0.42419732941521543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,2048,0.007675555845101674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,2560,0.011921777493423887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,10240,0.056892441378699414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,5120,0.010298666854699453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,3072,0.014041778114106921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,3584,0.016417778200573392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,256,0.005629333357016246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,768,0.00869333330127928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,128,0.020999110407299463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,4096,0.0144533332851198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,3072,0.0868613322575887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,768,0.0032071111102898917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,2560,0.006955555743641323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,2048,0.035679999325010515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,10240,0.023345778385798138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,2560,0.0074337778819931885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,512,0.003640000190999773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,16384,0.3667902151743571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,6144,0.04638488756285774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,3584,0.017431111799346078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,1024,0.029456890291637842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,768,0.002459555657373534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,7168,0.01148000028398302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,64,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,16384,1.4884701834784613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,12288,0.027965333726671007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,768,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,4096,0.008884444832801819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,6144,0.25090310308668345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,5120,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,3072,0.01032266683048672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,4096,0.02029511166943444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,7168,0.012338666452301873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,65536,0.3768746587965224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,16384,0.03314133485158285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,6144,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,2560,0.10181599855422974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,4096,0.005668444352017508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,12288,0.014139556222491793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,16384,0.07291289170583089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,8192,0.010672888822025724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,3072,0.009404444032245213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,2048,0.004490666505363253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,3584,0.00868977771864997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,128,0.0021217777911159727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,16384,0.005500444521506627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,256,0.006697777658700943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,768,0.011028444601429833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,6144,0.004937777916590373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,10240,0.029821332958009508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,2560,0.007242666350470648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,768,0.005630222045713001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,12288,0.07181066936916776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,7168,0.019239111079110038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,8192,0.011377777490350934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,10240,0.012169777519173093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,16384,0.017942221628295053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,128,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,3584,0.04031199879116482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,32,0.0019040000107553268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,768,0.015126221709781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,4096,0.05743555227915446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,16384,0.26874756813049316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,256,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,65536,0.08666488859388564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,2048,0.010274666878912184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,65536,0.07677955759896173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,65536,0.07591466771231757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,32,0.005650666852792104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,10240,0.015789333317014907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,8192,0.09494755665461223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,2560,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,256,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,6144,0.49840180079142254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,2560,0.008989333278603023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,1024,0.0041644444896115195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,6144,0.009755555954244401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,2048,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,2048,0.010168000227875179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,65536,1.1090372933281791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,7168,0.03958755400445726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,10240,0.0451528893576728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,1024,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,4096,0.00835111074977451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,768,0.0303457776705424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,2048,0.017235555582576327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,10240,0.14337066809336343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,7168,0.020654223031467862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,8192,0.025826666090223525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,10240,0.018014222383499146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,4096,0.02702666653527154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,128,0.0018320000833935207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,1536,0.02388711108101739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,10240,0.020854221449957952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,1024,0.019902222686343722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,6144,0.07099466853671603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,7168,0.028350220786200628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,12288,0.04311733444531759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,6144,0.009720000127951304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,3072,0.008464888566070134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,10240,0.03457244568400913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,7168,0.03343733482890659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,10240,0.012084444363911947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,8192,0.009344889058007134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,512,0.007287110719415877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,5120,0.012075555821259817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,8192,0.10415466626485188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,4096,0.06895466645558675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,16384,0.03258666727277968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,1024,0.00351200004418691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,2048,0.01162755572133594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,7168,0.01891644464598762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,128,0.0028248888750871024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,3584,0.018179555733998615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,3584,0.029331554969151814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,12288,0.45369868808322483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,3072,0.008956444760163626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,65536,0.06359289089838664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,7168,0.014399111270904541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,3584,0.013761777844693927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,2048,0.019215111931165058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,256,0.014442667365074158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,32,0.0018177777528762817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,12288,0.019913777709007263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,1536,0.007965332931942409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,7168,0.010716444088353051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,12288,0.018179555733998615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,7168,0.015172445111804538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,256,0.10665688912073772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,2048,0.008642666869693333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,1024,0.006304889089531369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,256,0.0022524444179402455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,6144,0.0488373339176178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,7168,0.03754577702946133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,1024,0.016177778442700703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,768,0.007300444775157505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,128,0.0031431110368834603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,8192,0.008278222547637092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,2048,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,32,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,3584,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,32,0.002187555655837059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,16384,0.34662487771775985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,3584,0.025527111358112756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,256,0.010046222143703038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,10240,0.11121155156029595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,6144,0.10871199766794841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,64,0.00509688879052798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,2560,0.013250666360060373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,10240,0.01608088943693373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,768,0.011021333436171213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,8192,0.049230221245023936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,512,0.0025457777082920074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,65536,0.07675821913613214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,2048,0.007817777494589487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,8192,0.031126221021016438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,256,0.027247111002604168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,2048,0.004222222086456087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,10240,0.010073777702119615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,2048,0.007598222129874759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,1024,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,65536,0.07089955276913114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,5120,0.02016088863213857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,8192,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,10240,0.018608889646000333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,3584,0.01109333336353302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,65536,0.12877244419521755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,64,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,6144,0.03010666701528761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,512,0.014083556003040738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,2048,0.11496799521976048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,2048,0.01683111157682207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,5120,0.025277333127127752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,1024,0.009309333231714036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,768,0.016144000821643405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,2560,0.007249777515729268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,10240,0.07411733600828382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,128,0.012379555238617791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,32,0.002485333424475458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,65536,0.0641626649432712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,65536,0.7064080238342285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,10240,0.050127112203174166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,128,0.006919999917348226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,6144,0.0682800014813741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,5120,0.014442667365074158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,3584,0.008094222181373173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,16384,0.0795253316561381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,8192,1.432728025648329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,8192,0.12225245104895698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,1024,0.014207111464606391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,1024,0.006056889063782162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,5120,0.007599111232492659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,65536,0.8188595771789551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,12288,0.018238221605618794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,6144,0.013078221844302284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,32,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,7168,0.014356444279352823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,2560,0.0068915556702348925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,1536,0.0335440006521013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,10240,0.005623111294375525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,7168,0.042640888028674655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,1536,0.008119111259778341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,2048,0.011355555719799466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,4096,0.03090133269627889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,4096,0.016050666570663452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,1024,0.01277955538696713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,64,0.0018817778262827133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,3072,0.005642666584915585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,128,0.0018835556175973679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,128,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,4096,0.17004711098141143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,512,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,6144,0.01927288870016734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,8192,0.02269422180122799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,256,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,5120,0.01440711153878106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,4096,0.008106666306654612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,2048,0.04095822241571214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,512,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,256,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,6144,0.018548443913459778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,2560,0.011424000064531961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,5120,0.009205333060688442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,6144,0.027245332797368366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,2048,0.02808977829085456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,256,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,4096,0.011711111499203576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,128,0.010061333576838175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,65536,0.07107644610934787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,8192,0.018939556346999276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,4096,0.10111822022332086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,256,0.0025031111306614345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,16384,0.016167999969588388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,12288,0.05385155479113261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,4096,0.0240639994541804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,2560,0.04790311058362325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,6144,0.024273778001467388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,8192,0.026478222674793665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,4096,0.020077332854270935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,16384,0.01348355578051673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,2560,0.013780444032616086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,6144,0.016494222813182406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,12288,0.012071111136012606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,4096,0.020954666866196528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,768,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,5120,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,512,0.012853333519564735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,1536,0.004848888764778773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,6144,0.21226578288608125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,16384,0.03013155526585049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,2048,0.01166488892502255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,256,0.007179555793603261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,128,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,6144,0.008983111215962304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,32,0.002560888934466574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,10240,0.042746666404936046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,65536,0.11963733037312825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,64,0.008879111044936711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,5120,0.010841777755154504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,512,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,1024,0.0042204442951414315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,512,0.00424799985355801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,10240,0.053314665953318276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,2560,0.006680000159475539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,6144,0.013177778157922955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,2560,0.011504000259770287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,128,0.01718133356836107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,3584,0.03521777855025397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,3584,0.014115555418862237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,6144,0.009673777553770278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,65536,0.13367999924553767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,5120,0.4921101993984646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,3072,0.00888888869020674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,256,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,5120,0.030309332741631403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,768,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,2048,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,2048,0.013767110804716745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,1536,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,1536,0.006614222294754452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,3072,0.01444355481200748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,65536,0.20123999648623994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,2048,0.01513155632548862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,16384,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,2048,0.013420444395807056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,7168,0.24542223082648384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,8192,0.01773955590195126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,12288,0.027183110515276592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,2560,0.013884444203641681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,5120,0.033964445193608604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,2048,0.021754667162895203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,65536,0.131294223997328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,7168,0.04907377892070346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,2560,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,10240,0.03098755412631565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,2560,0.013747555514176687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,1536,0.01071911139620675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,3584,0.018372444642914664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,3072,0.0366284449895223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,12288,0.031077331966824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,2560,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,5120,0.03846755623817444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,4096,0.017329777280489605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,65536,0.046670221620135836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,512,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,128,0.00563733321097162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,4096,0.008264888491895463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,8192,0.03974577784538269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,3584,0.09888799985249837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,128,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,1536,0.004695999953481886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,1536,0.020563556088341605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,2048,0.01814133259985182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,4096,0.016880000631014507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,32,0.001779555653532346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,12288,0.025592888394991558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,16384,0.055163555675082736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,10240,0.07384710841708712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,256,0.015216888652907478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,256,0.030277334981494482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,10240,0.04984622200330099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,512,0.013788444300492605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,1536,0.0037226668662495087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,16384,0.035944888989130654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,4096,0.01687022215790219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,3584,0.008303999900817871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,3584,0.011271110839313932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,4096,0.008811555802822113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,5120,0.014375999569892883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,2560,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,768,0.09538133276833428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,7168,0.012110222544935016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,6144,0.012224888636006249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,128,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,10240,0.056427558263142906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,5120,0.009357333183288574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,8192,0.04834666517045763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,4096,0.009659555223253038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,8192,0.02159822152720557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,768,0.010582221878899468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,3584,0.007972444097201029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,10240,0.02336799932850732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,10240,0.008708444734414419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,512,0.006680888848172293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,12288,0.06250488758087158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,512,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,4096,0.033379554748535156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,2560,0.01220355596807268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,2048,0.004862222406599256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,768,0.00281866660548581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,1536,0.004881777697139316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,65536,0.19550666544172499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,64,0.0018231111268202465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,1024,0.011738667057620155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,128,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,12288,0.04074577821625604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,3072,0.021385777327749465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,65536,0.06050399939219157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,1536,0.0063031112982167145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,5120,0.009005332986513773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,768,0.030807110998365615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,4096,0.009011555876996782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,1024,0.015549333559142219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,12288,0.03044355577892727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,16384,0.045810666349199086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,4096,0.012362666428089142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,7168,0.013387555877367655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,256,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,512,0.0025742221623659134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,8192,0.015192889504962497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,10240,0.010040889183680216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,5120,0.008662222160233391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,2048,0.0075635554061995606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,3584,0.010649777948856354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,12288,0.05071466830041674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,6144,0.02033866610791948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,10240,0.02462311089038849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,65536,0.08358666631910537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,4096,0.023334221707450017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,10240,0.044016887744267784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,4096,0.015456888410780164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,1536,0.014095111025704278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,1536,0.028620445066028174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,256,0.002465777720014254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,65536,0.06537510951360066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,2560,0.017138666576809354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,8192,0.2763964335123698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,4096,0.0718382199605306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,5120,0.03084889054298401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,1536,0.006351111249791251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,3584,0.009023110899660323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,3072,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,5120,0.026771555344263714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,512,0.013849777479966482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,65536,0.18685955471462676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,7168,0.023711111810472276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,3072,0.012183999849690331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,6144,0.012826667063766055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,768,0.07519733243518405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,1024,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,512,0.005554666535721884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,7168,0.01753422286775377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,128,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,65536,0.11193422476450603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,4096,0.011340444286664328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,12288,0.03356533249219259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,128,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,10240,0.02091911103990343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,1536,0.01818933255142636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,7168,0.058796445528666176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,1024,0.013788444300492605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,2048,0.03524355424775018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,7168,0.013831111292044321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,65536,0.15047910478379992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,256,0.014936889211336771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,7168,0.010409778190983666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,64,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,8192,0.004970666848950916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,5120,0.017864000466134813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,65536,0.0650728874736362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,768,0.014762666490342883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,2048,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,512,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,12288,0.01941422290272183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,3072,0.016527111331621807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,8192,0.02997777859369914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,2560,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,65536,0.05333066648907132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,7168,0.010036444498433007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,5120,0.040576001008351646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,512,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,3072,0.008907555705971187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,10240,0.018228444788191054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,10240,0.012141333685980903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,3584,0.01684977776474423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,8192,0.03099200129508972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,128,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,512,0.06327911218007405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,3072,0.01685066686736213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,5120,0.016189333465364244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,16384,0.01578044394652049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,512,0.0052266667286554975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,256,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,5120,0.05889155467351278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,64,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,2048,0.012735111018021902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,5120,0.1473697821299235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,5120,0.037003556887308754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,4096,0.007648888561460707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,7168,0.011005333728260465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,256,0.0052915554907586836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,4096,0.0226248883538776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,5120,0.01682933337158627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,256,0.005978666659858491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,5120,0.08106933037439983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,10240,0.03365244468053182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,16384,0.019690665933820937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,4096,0.009275555610656738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,256,0.002340444467133946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,64,0.002514666567246119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,6144,0.024893333514531452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,768,0.007280888656775157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,3584,0.010002666877375709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,2048,0.0229751103454166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,2560,0.03291822142071194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,2048,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,6144,0.02757155564096239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,768,0.005640888793600931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,1536,0.01579466627703773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,64,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,128,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,512,0.002524444419476721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,3584,0.018752889500723947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,12288,0.18722132841746011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,6144,0.01312355531586541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,128,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,3584,0.01889333294497596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,16384,0.05440622236993578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,2560,0.014689778288205465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,8192,0.007712889048788283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,6144,0.01073333372672399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,2560,0.0075457774930530125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,6144,0.010052444206343757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,2048,0.022250667214393616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,7168,0.015736889508035447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,6144,0.009639999932712978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,1024,0.01605955594115787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,64,0.005610666755172942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,512,0.01684533390733931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,128,0.002163555473089218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,8192,0.010701333483060202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,8192,0.009515555368529426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,16384,0.0699768861134847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,10240,0.025411556164423626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,8192,0.02749600013097127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,7168,0.04229866796069675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,7168,0.014469333820872836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,1536,0.011896000140243106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,1536,0.016879111528396606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,3072,0.0982053346104092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,8192,0.024749333659807842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,5120,0.01311466677321328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,6144,0.01886577738655938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,128,0.01443199978934394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,4096,0.008010666403505538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,128,0.005251555393139522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,3584,0.008658666577604082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,768,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,16384,0.02053511142730713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,5120,0.033089776833852134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,256,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,3584,0.007582222421964009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,8192,0.015269334117571512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,128,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,1024,0.00349866681628757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,512,0.021951999929216173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,4096,0.011725333001878528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,2048,0.026963555150561865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,256,0.03273511264059279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,16384,0.09849333100848728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,16384,0.21978488233354357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,2048,0.02128444446457757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,12288,0.011055111057228513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,3584,0.008801778157552084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,768,0.0052284445199701525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,6144,0.015096000499195524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,8192,0.04097244474622939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,768,0.01811733345190684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,128,0.014500444134076437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,16384,0.01592177814907498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,10240,0.02439466615517934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,2048,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,10240,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,768,0.008001777860853408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,65536,0.13957866032918295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,256,0.015450666348139444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,512,0.002499555548032125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,7168,0.01035644445154402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,256,0.01001422190003925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,10240,0.024380443824662104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,6144,0.02804622385236952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,8192,0.07144888904359606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,65536,0.10731644762886895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,2048,0.01751733322938283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,4096,0.1127119991514418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,128,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,3584,0.008719999757077958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,5120,0.01240533341964086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,128,0.013405332962671915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,2560,0.01680088871055179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,65536,0.08494044674767388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,8192,0.006303999986913469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,32,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,3584,0.014594667487674289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,8192,0.012353777885437012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,768,0.005630222045713001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,1024,0.009696889254781935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,512,0.005371555685997009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,65536,0.05049066742261251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,7168,0.15349422560797796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,768,0.01758666667673323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,12288,0.013428444663683573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,2560,0.007236444287829929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,7168,0.018639999959203932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,2048,0.0251519994603263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,16384,0.016163556112183463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,7168,0.02948533495267232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,16384,0.09192533625496759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,256,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,3584,0.020871111088328894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,8192,0.011042666931947073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,16384,0.005737777799367905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,12288,0.03568799959288703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,32,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,512,0.002504000026318762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,7168,0.020922667450375024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,16384,0.016550223032633465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,3072,0.011362666885058085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,12288,0.6852684550815158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,64,0.010517333116796283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,65536,0.14161422517564562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,65536,0.08673333459430271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,3584,0.008319999608728621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,6144,0.012019555601808759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,5120,0.011366222467687396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,256,0.005694222119119432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,1536,0.01329422245422999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,6144,0.019648000597953796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,8192,0.08093866374757555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,5120,0.01927288870016734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,4096,0.008949333594905006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,3584,0.006212444355090459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,512,0.013448889056841532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,4096,0.010399999717871347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,65536,0.046577778127458364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,16384,0.05093066559897529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,2048,0.009295110901196798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,3584,0.00872622181971868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,8192,0.011694221860832639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,2048,0.018191999859280057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,2560,0.015882665912310284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,3072,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,7168,0.4777013460795085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,6144,0.00963466697269016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,10240,0.012439111040698158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,768,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,1536,0.006248888870080312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,1024,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,8192,0.024989333417680528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,2560,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,10240,0.011741333537631564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,8192,0.01204622205760744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,5120,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,2048,0.005190222213665645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,65536,0.12399289343092178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,1024,0.00832088871134652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,2560,0.015397333436542086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,32,0.0018240000224775737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,65536,0.16668355464935303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,256,0.011197333534558615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,2560,0.0069066666894488865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,4096,0.07537155681186251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,10240,0.01480977733929952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,2048,0.004524444540341695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,3584,0.1822284460067749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,4096,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,6144,0.05794488721423679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,2560,0.009315555294354757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,32,0.0028719999310043124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,128,0.003479111111826367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,8192,0.046684443950653076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,768,0.0151555554734336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,768,0.04626933402485318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,3584,0.007667555577225155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,6144,0.011400889191362592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,16384,0.030160887373818293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,3072,0.0934106641345554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,10240,0.029205332199732464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,128,0.0064151109092765385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,1536,0.007299555672539606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,128,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,768,0.004520888957712385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,3072,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,1024,0.0030826667530669104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,768,0.03713599840799967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,1536,0.08299822277492948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,12288,0.05236000153753492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,256,0.0052844443255000645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,5120,0.010703999963071613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,6144,0.012746666868527731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,10240,0.024420445164044697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,2560,0.0074444446298811175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,4096,0.011032000184059143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,16384,0.018925334016482036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,8192,0.014184888866212634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,64,0.0038417776425679526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,3584,0.012440000143316058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,10240,0.03458044595188565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,256,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,5120,0.009573332965373993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,10240,0.09414666891098022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,4096,0.00815466625822915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,7168,0.04395555456479391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,1536,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,3584,0.00423200014564726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,10240,0.6694666544596354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,1024,0.003552000141806073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,7168,0.015304888288180033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,3072,0.00792266676823298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,65536,0.36569955613878036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,2048,0.009720889230569204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,512,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,1536,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,512,0.002534222271707323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,128,0.0151235560576121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,16384,0.04432444439993965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,768,0.0185253338681327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,1024,0.009328000247478485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,7168,0.029355555772781372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,7168,0.031686223215527005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,10240,0.03185155656602647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,6144,0.011634666886594562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,5120,0.04699466625849406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,3072,0.02239911092652215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,2048,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,2560,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,10240,0.06258755260043673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,12288,0.014820444915029736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,3072,0.004537777768241035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,1536,0.030685332086351182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,6144,0.016208888755904306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,2560,0.00941777808798684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,6144,0.009015999734401703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,1024,0.0053324442770746015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,64,0.002817777709828483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,2048,0.006555555595291986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,65536,0.0831164452764723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,6144,0.004922666483455234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,64,0.004521777646409141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,3584,0.01637066735161675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,4096,0.048143111997180514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,10240,0.02681511143843333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,4096,0.04227288895183139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,10240,0.029570665624406602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,12288,0.034367998441060386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,10240,0.02706577711635166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,2048,0.010517333116796283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,16384,0.019310222731696237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,10240,0.9425652821858724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,1024,0.013573333621025085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,16384,0.03908888830078973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,768,0.009667555491129557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,768,0.07233244180679321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,256,0.01651022169325087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,1536,0.009300444689061906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,1024,0.007631111476156447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,12288,0.02147377696302202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,1536,0.00488800017370118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,5120,0.0433919992711809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,256,0.0021706666383478376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,3072,0.03220622075928582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,5120,0.03280266788270738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,12288,0.020075556304719713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,256,0.005669333454635408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,16384,0.012321777641773224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,256,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,2560,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,2048,0.004582222137186262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,12288,0.013741333451535968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,1536,0.030128889613681372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,128,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,3584,0.14126400152842203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,4096,0.008525333470768398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,256,0.011439999772442712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,2560,0.0042168887125121225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,12288,0.049439999792310924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,1024,0.07287555270724826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,512,0.03389777739842733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,16384,0.2288231054941813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,128,0.01477599971824222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,128,0.006636444479227066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,1536,0.0038337777886125776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,12288,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,3072,0.01484444406297472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,512,0.002943111169669363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,3584,0.027056000298923914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,2048,0.030047112041049536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,1536,0.054343998432159424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,12288,0.06565422481960721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,256,0.005884444548024072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,16384,0.06826311349868774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,10240,0.023431110713216994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,128,0.002477333363559511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,3072,0.05563555823432075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,128,0.01514933341079288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,512,0.013415999710559845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,1024,0.028068443139394123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,256,0.01682222220632765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,1536,0.015847111741701763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,10240,0.02690222197108799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,8192,0.0413955549399058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,65536,0.9612568749321831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,2560,0.10038933489057754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,16384,0.058669335312313504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,12288,0.029676444000667993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,1024,0.03867377837498983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,10240,0.0325582226117452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,5120,0.2010115517510308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,10240,0.4462622006734212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,16384,0.03357333276006911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,128,0.01683111157682207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,1536,0.015112888481881885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,10240,0.08456177843941583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,3072,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,7168,0.12722222010294595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,6144,0.11266133520338271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,768,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,8192,0.04302577839957344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,1536,0.008656889200210571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,7168,0.22712445259094238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,2560,0.029746668206320867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,2560,0.013847110999955071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,2048,0.004607999904288186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,16384,0.028571556011835735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,128,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,6144,0.009664889011118147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,1536,0.026963555150561865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,1024,0.026603556341595117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,3584,0.009552000297440423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,1024,0.005974222388532426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,768,0.007291555404663086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,7168,0.030237333642111883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,256,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,12288,0.065138664510515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,64,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,2560,0.0075777777367168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,65536,0.06937511099709405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,12288,0.02539644473128849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,3584,0.009988444546858469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,6144,0.01997866729895274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,4096,0.007661333514584436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,12288,0.6290204260084364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,10240,0.02094399929046631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,10240,0.012039999994966718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,7168,0.015102222561836243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,12288,0.03857422206136916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,10240,0.012057777908113269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,65536,0.11907377507951523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,12288,0.03124177787039015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,5120,0.008619555168681675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,2560,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,4096,0.03029511041111416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,128,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,768,0.014294221997261047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,16384,0.04831288920508491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,7168,0.027414222558339436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,256,0.005402666827042897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,10240,0.017484444710943434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,512,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,128,0.0018320000833935207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,7168,0.041430221663581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,3072,0.004237333519591225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,6144,0.01665422154797448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,64,0.0024782222592168385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,6144,0.030202666918436687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,3072,0.0476435555352105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,65536,0.05776800049675835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,512,0.0026382222357723448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,128,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,7168,0.013457777599493662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,256,0.01615111033121745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,16384,0.18244622813330758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,1536,0.02825777729352315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,4096,0.047016001409954496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,128,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,32,0.017550221747822232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,7168,0.0099982221921285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,512,0.002536000063021978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,3072,0.007328000333574083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,3072,0.018180444836616516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,2560,0.017885333961910672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,7168,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,4096,0.01236355553070704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,16384,0.027066666218969557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,4096,0.020326221982638042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,8192,0.05039199855592516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,3584,0.02955733405219184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,65536,0.048615998691982694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,5120,0.01734044485621982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,1024,0.012378666963842181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,2048,0.008715555899673039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,1024,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,10240,0.4308791160583496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,512,0.007224000162548489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,4096,0.011857777833938599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,2560,0.007844444778230455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,1024,0.006227555374304454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,768,0.016128889388508268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,6144,0.02067466742462582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,128,0.024487111303541396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,5120,0.057201776239607066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,16384,0.021382222572962444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,128,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,12288,0.0905182229148017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,4096,0.1395217842525906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,4096,0.16045688258277044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,3072,0.0135057775510682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,6144,0.023815110325813293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,8192,0.06005777915318807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,7168,0.03609066539340549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,64,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,256,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,2560,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,768,0.004084444294373195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,16384,0.02607288956642151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,12288,0.06048355499903361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,2560,0.05817688835991753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,3584,0.008852444589138031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,3584,0.004896000027656555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,16384,0.2010320027669271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,1024,0.02968533171547784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,4096,0.014800000521871777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,3072,0.010760889285140567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,1024,0.002838222309947014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,768,0.017575111654069688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,8192,0.03206933206982083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,7168,0.008443555898136562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,1024,0.015474667151769003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,12288,0.017866666118303936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,128,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,7168,0.011407111254003314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,12288,0.013426666458447775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,256,0.005686222265164058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,4096,0.04028266668319702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,10240,0.16157866848839655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,12288,0.04622044497066074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,512,0.012268444730175866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,3584,0.00830488900343577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,1024,0.007535999847782983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,32,0.0021422221842739317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,7168,0.02810488806830512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,7168,0.027759111589855615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,5120,0.00452888881166776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,4096,0.012329777909649743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,768,0.039460443788104586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,2560,0.016158221496476066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,128,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,768,0.015365334020720588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,64,0.0017964444640609955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,16384,0.28168800142076283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,10240,0.023606222536828782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,8192,0.034767998589409724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,2048,0.013849777479966482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,768,0.08178489075766669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,3584,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,10240,0.015511110424995422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,12288,0.028549333413441975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,1024,0.014101333088344999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,16384,0.016523554921150208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,256,0.009823111196359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,6144,0.013642667068375481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,6144,0.009695111049546136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,12288,0.0199617776605818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,16384,0.006030222194062338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,10240,0.027237332529491846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,7168,0.06982755661010742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,7168,0.021729777256647747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,12288,0.01889955500761668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,65536,0.06387022468778822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,2048,0.006649777707126405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,2560,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,512,0.014114666316244336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,1024,0.007738666401969061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,16384,0.021944888763957556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,16384,0.019205333458052743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,7168,0.11867554982503255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,768,0.003858666867017746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,4096,0.004536889079544279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,6144,0.007439110842016008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,65536,1.6263537936740453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,128,0.014447111222479077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,512,0.0083155557513237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,2560,0.009534222384293875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,2048,0.009120889008045197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,8192,0.012051555845472546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,64,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,768,0.018554665976100497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,2560,0.005249777601824866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,12288,0.03326755430963304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,2560,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,1024,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,128,0.01477866702609592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,16384,0.02623111175166236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,32,0.0021377777059872947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,3072,0.0388995541466607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,768,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,1536,0.023327999644809302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,32,0.0025235555238193935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,2560,0.006719999843173557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,128,0.011040888726711273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,10240,0.01309511148267322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,5120,0.008648888932334052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,10240,0.01565511193540361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,65536,3.2907049391004772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,65536,0.14808889230092367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,128,0.010409778190983666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,512,0.009333333207501305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,8192,0.016502221425374348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,16384,0.26299821005927193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,7168,0.013352000051074557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,12288,0.15235200193193224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,1024,0.015121777852376303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,6144,0.058873779243893094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,65536,0.06021777788798014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,64,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,2048,0.023447111248970032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,65536,0.08474577797783746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,3584,0.016099555624855887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,2560,0.06287911203172472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,65536,0.6158097585042318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,16384,0.06425511174731784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,1536,0.008664000365469191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,1536,0.07979910903506808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,3584,0.0074240002367231585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,65536,0.06867644521925184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,4096,0.007630222373538547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,3072,0.018339556124475267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,2048,0.011544889046086205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,7168,0.0042053332759274375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,3584,0.06264177958170573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,3584,0.00942311104800966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,3072,0.013456000222100152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,4096,0.030896888838873968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,768,0.0737288859155443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,1536,0.003571555432346132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,3072,0.016568000117937725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,5120,0.025193777349260118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,32,0.013775111072593264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,7168,0.01925244430700938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,16384,0.06319377819697063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,7168,0.010070222119490305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,1536,0.0038399998512532976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,2048,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,65536,0.1848897801505195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,1024,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,7168,0.036312888065973915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,7168,0.022441777918073866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,768,0.0028719999310043124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,2560,0.02351733379893833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,3584,0.1424106624391344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,8192,0.013850666582584381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,2048,0.005668444352017508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,12288,0.06323377953635322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,65536,0.22706132464938691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,3072,0.056871109538608126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,256,0.019855111837387085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,8192,0.018216888109842937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,8192,0.22039288944668242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,12288,0.01965688831276364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,2560,0.013440888788965015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,16384,12.68995327419705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,10240,0.04904266529613071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,256,0.006309333360857434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,12288,0.06480089161131117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,2048,0.012083555261294046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,7168,0.02424977719783783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,3584,0.01717511150572035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,256,0.014687110980351767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,256,0.013077333569526672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,16384,0.025409777959187824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,5120,0.11029777924219768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,32,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,512,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,128,0.014448000325096978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,8192,0.01038755559258991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,12288,0.03144977821244134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,7168,0.010727110836240979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,65536,1.2237520217895508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,5120,0.006907555378145642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,1024,0.0063582224150498705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,512,0.0032631111227803757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,3584,0.15614222155676946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,65536,0.011012444065676795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,256,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,128,0.005392888767851724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,2048,0.01035377797153261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,5120,0.008960888617568547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,5120,0.05747199720806546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,16384,0.02163644466135237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,128,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,32,0.0038702223036024305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,5120,0.010755555497275459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,32,0.0018639999131361644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,16384,0.025081778566042583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,2560,0.00630666646692488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,2048,0.013021333350075616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,7168,0.010355555348926121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,10240,0.01809955636660258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,6144,0.009333333207501305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,256,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,12288,0.0188080006175571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,2560,0.01349688900841607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,64,0.001954666649301847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,128,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,6144,0.02532622218132019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,768,0.01696888936890496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,2048,0.0069520001610120135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,512,0.01015733347998725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,3072,0.015556444724400839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,12288,0.018171555466122098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,1536,0.004926222066084544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,2048,0.006001777946949005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,6144,0.03308799862861633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,768,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,8192,0.15051200654771593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,1536,0.008671999805503422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,8192,0.013712888790501488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,12288,0.1398871077431573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,65536,0.019891555110613506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,2048,0.09778666496276855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,3584,0.016516443755891588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,128,0.003140444349911478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,7168,0.04615911179118686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,16384,0.1394995583428277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,512,0.007319110963079665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,64,0.0052764444715446895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,10240,0.008910222185982598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,128,0.015834665960735746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,5120,0.027122666438420612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,3584,0.014089778065681458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,3072,0.02348711093266805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,8192,0.40596710311041934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,65536,0.31437865893046063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,6144,0.03578311204910278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,64,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,6144,0.01942755613062117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,6144,0.032327112224366926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,768,0.017161778277821012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,10240,0.021923555268181696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,64,0.006562666760550604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,1024,0.01924622224436866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,512,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,64,0.002524444419476721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,1536,0.006330666856633292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,4096,0.18051555421617296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,3584,0.013915555344687568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,12288,0.008799999952316284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,128,0.001960888918903139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,4096,0.020601777566803824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,1536,0.027867555618286133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,12288,0.11772888236575657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,3584,0.010371555884679159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,3072,0.013551111022631327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,6144,0.009113777842786577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,768,0.018315555320845712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,16384,0.02214755614598592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,64,0.006302222195598815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,6144,0.019703111714786954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,64,0.006317333214812809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,10240,0.12685510847303602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,10240,0.022758222288555566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,6144,0.5230026774936253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,10240,0.03810933232307434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,512,0.0032542221662071017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,10240,0.03663911090956794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,32,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,32,0.007689778175618913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,10240,0.03619911273320516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,4096,0.008651555412345463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,256,0.009703111317422655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,2048,0.007967111137178209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,768,0.021341333786646526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,64,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,8192,0.2115759982003106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,1024,0.013759111364682516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,128,0.007968000239796108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,1024,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,6144,4.607326931423612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,256,0.003181333343187968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,16384,0.05375644564628601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,1024,0.007648888561460707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,7168,0.019263111882739596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,7168,0.1033173335923089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,3072,0.007383999725182851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,512,0.018532445033391316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,64,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,8192,0.023042667243215773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,7168,0.019681778219011094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,3584,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,2048,0.0069368887278768755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,3584,0.013732444081041547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,2560,0.04177777634726631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,512,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,3584,0.3826133410135905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,32,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,7168,0.026365333133273657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,6144,0.023362666368484497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,2048,0.12477955553266738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,16384,0.03534400131967332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,768,0.01646577815214793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,4096,0.00795111142926746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,3584,0.02961777647336324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,4096,0.008288000192907121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,10240,0.0262080000506507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,8192,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,3072,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,5120,0.011680888632933298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,256,0.0052675555149714155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,1536,0.01550044450494978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,32,0.0018035556293196147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,1024,0.026425777210129633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,6144,0.024811555941899616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,6144,0.00981155534585317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,2560,0.011708444191349877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,1024,0.009284444153308868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,12288,0.02530133393075731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,4096,0.011232888533009423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,3072,0.008009778128729926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,64,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,256,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,3072,0.007532444265153673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,2560,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,65536,0.05659288830227322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,2560,0.01886577738655938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,8192,0.02129688858985901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,8192,0.0311173333062066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,2560,0.019468444916937087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,4096,0.017144888639450073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,128,0.001842666624320878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,7168,0.14184266991085476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,768,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,2048,0.00453333349691497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,16384,0.02753777801990509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,12288,0.17561422453986275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,65536,1.6521547105577257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,4096,0.01110311100880305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,12288,0.022664889693260193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,65536,0.10406044456693862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,32,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,2560,0.04343466626273262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,12288,0.04394488864474826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,1024,0.017865777015686035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,2560,0.08639999892976548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,16384,0.24157688352796766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,256,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,12288,0.01310400002532535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,6144,0.01035911093155543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,3584,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,3072,0.023200000325838726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,12288,0.032088001569112144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,512,0.005416888743638992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,768,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,256,0.004944889081848992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,2560,0.03421599997414483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,12288,0.03166133496496413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,768,0.009391110804345872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,1536,0.022321777211295232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,6144,0.01036266651418474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,10240,0.019850666324297588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,6144,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,7168,0.015060444672902426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,128,0.013779555757840475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,2560,0.007449777589903937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,65536,0.12105422549777561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,4096,0.024682665864626568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,1024,0.04082399937841628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,12288,0.060319999853769936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,2048,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,256,0.0034062222888072333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,128,0.006527110934257507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,12288,0.0444764428668552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,1536,0.01826133330663045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,8192,0.01069777790043089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,128,0.014498665928840637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,5120,0.027628444963031348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,5120,0.017311111092567444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,256,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,1024,0.0028391112056043413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,2560,0.008206222620275285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,65536,0.08620978064007229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,512,0.00630044440428416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,6144,0.016175111134847004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,65536,2.076113806830512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,3584,0.01923288901646932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,10240,0.019567999574873183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,1536,0.009399111072222391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,32,0.008654221892356873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,12288,0.031122667921913996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,4096,0.026749332745869953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,3072,0.028335111008750066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,768,0.019736889335844252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,3584,0.7845537927415637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,2048,0.00665866666369968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,8192,0.027313777142100867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,64,0.001822222231162919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,65536,0.0574551092253791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,3584,0.010683555569913654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,32,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,2048,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,768,0.005761777775155173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,64,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,12288,0.01687111126052009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,1024,0.006298666612969504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,10240,2.2342169019911027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,12288,2.3592987060546875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,1536,0.024698666400379602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,512,0.021627555290857952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,512,0.021513778302404616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,512,0.06108355522155762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,128,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,16384,0.022129777404997084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,3072,0.007609777980380588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,6144,0.0313991109530131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,2560,0.04968711071544223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,2048,0.013405332962671915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,768,0.018539556198649935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,768,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,8192,0.028428445259730022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,3584,0.02717155549261305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,3584,0.02402666707833608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,3584,0.014070222775141398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,512,0.007252444823582967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,3584,0.0073831114504072405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,12288,0.05379466546906365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,3072,0.014833778142929077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,1536,0.024452444579866197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,3584,0.022271111607551575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,6144,0.018630221486091614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,1536,0.016378665963808697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,512,0.0028479999552170434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,16384,0.2352213329739041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,65536,0.0754257771703932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,1024,0.016179554992251925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,1024,0.005873777800136142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,256,0.0151582227812873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,6144,0.05216622352600098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,256,0.03183822168244256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,5120,0.2768782244788276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,3584,0.007631999750932057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,512,0.02846133377816942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,6144,0.04774489005406698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,256,0.0022355556074115965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,768,0.017535110314687092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,8192,0.012061333490742577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,64,0.007387555307812161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,10240,0.24130845069885254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,8192,0.06709600157207914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,768,0.007631999750932057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,32,0.0018008889423476325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,32,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,10240,0.06913333468967013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,6144,0.12450310919019912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,16384,0.3649928834703233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,8192,0.022322666313913133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,512,0.014279110564125909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,12288,0.1088106698460049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,6144,0.012392000191741519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,768,0.005517333332035277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,256,0.014498665928840637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,2560,0.016227554943826463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,10240,0.00812888890504837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,64,0.006268444574541516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,3072,0.030393779277801514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,2048,0.004262222184075249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,65536,0.03717689050568475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,65536,0.07070133421156141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,512,0.013819555441538492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,16384,0.08260355393091838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,4096,0.21083822515275744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,16384,0.061186664634280734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,1536,0.008406221866607666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,5120,0.02537333303027683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,512,0.01620533400111728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,1536,0.00665511108107037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,16384,0.037793778710895114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,1536,0.003563555578390757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,65536,0.06725866927040948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,6144,0.010325333310498131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,64,0.0094035557574696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,2048,0.00829777783817715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,4096,0.02889510989189148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,2048,0.01090666651725769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,65536,0.07686666647593181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,3072,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,128,0.01443288889196184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,64,0.0021226666867733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,16384,0.04042133357789781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,512,0.012768000364303589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,12288,0.04663911130693224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,768,0.014119111829333834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,768,0.01884622209601932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,8192,0.036213331752353244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,10240,0.0964897804790073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,16384,0.06435288985570271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,6144,0.013791999883121915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,2048,0.00463733325401942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,1536,0.13992533418867323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,5120,0.009524444739023844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,4096,0.009909333454238044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,512,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,7168,0.025232889586024817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,16384,0.005883555445406172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,65536,0.0069155556460221606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,10240,0.022794667217466567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,5120,0.025507556067572698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,3072,0.005723555468850666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,128,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,16384,0.05396533343527052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,16384,0.012326222327020435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,5120,0.027427555786238775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,5120,0.015088889333936902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,4096,0.05572622352176242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,1536,0.015391111373901367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,512,0.01444533301724328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,512,0.005623111294375525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,16384,0.0395066671901279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,12288,0.03113244308365716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,12288,0.017875555488798354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,2048,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,16384,2.4885662926567926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,3584,0.009424000150627559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,128,0.01719199948840671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,16384,0.05816177527109782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,8192,0.15736533535851374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,12288,0.33591734038458926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,256,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,8192,0.070196443133884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,3584,0.02642755541536543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,6144,0.3427386548784044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,65536,0.6386639806959364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,8192,0.024432000186708238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,768,0.01754488878779941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,12288,0.2482079929775662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,2048,0.018441778090265062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,12288,0.050716443194283374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,2560,0.013447110851605734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,1536,0.006820444431569841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,5120,0.02915999955601162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,2048,0.024751999311976965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,64,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,1024,0.014310222533014087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,65536,0.1485955582724677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,512,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,12288,0.019105777144432068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,2560,0.004567999806669024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,256,0.016185777054892648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,16384,0.028155555327733357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,65536,0.08529333273569743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,3072,0.007442666424645319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,6144,0.023183999790085688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,256,0.0022248888595236675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,5120,0.009692444569534725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,65536,0.15531467066870794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,16384,0.10067199998431736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,3072,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,65536,0.12065066231621636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,1536,0.010644444160991244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,10240,0.011720889144473605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,1536,0.016162667009565566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,12288,0.008968888885445064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,5120,0.012290666500727335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,10240,0.016176889340082806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,65536,0.07912711302439372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,65536,0.768155574798584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,7168,0.015112888481881885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,3072,0.013545778062608508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,1536,0.02490844494766659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,256,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,128,0.005579555614127054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,1024,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,12288,0.017940445078743827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,1024,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,12288,0.034535109996795654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,4096,0.007780444290902879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,16384,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,4096,0.033886220720079206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,65536,0.07463466458850436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,5120,0.022946667340066697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,256,0.011019555230935415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,768,0.017702221870422363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,768,0.020202666521072388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,10240,0.02271555529700385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,128,0.005943999936183293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,512,0.015125332607163323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,2560,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,32,0.001847111102607515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,8192,0.013316444224781461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,128,0.00563466673096021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,8192,0.036237332555982806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,3584,0.007316444483068254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,3584,0.016472889317406546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,3072,0.02293866707219018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,64,0.003531555748648114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,256,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,32,0.005708444449636672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,6144,0.009017777939637503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,16384,0.02124533388349745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,6144,0.013003555436929068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,768,0.016757332616382174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,3584,0.008007999923494127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,1536,0.004907555464241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,2048,0.02597155504756504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,768,0.007739555504586961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,3072,0.2622951136695014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,3072,0.03920711080233256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,32,0.05818133221732246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,3072,0.008964444200197855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,12288,0.048024889495637685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,128,0.0021333334346612296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,4096,0.018572444717089336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,65536,0.05169511172506544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,512,0.006241777704821692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,512,0.02697333362367418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,32,0.002493333278430833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,2560,0.015164444843928019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,10240,0.026687999566396076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,1536,0.0063075555695427795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,1024,0.01585155559910668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,512,0.002524444419476721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,16384,0.04280799958440992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,6144,0.14652089277903238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,16384,0.015819556183285184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,65536,0.3166968822479248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,768,0.04847200049294365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,5120,0.1955102284749349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,4096,0.010207110808955299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,2048,0.005714666512277391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,4096,0.011380444798204633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,5120,0.5188070933024088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,3072,0.017807111144065857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,256,0.015576000014940897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,8192,0.03658399979273478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,12288,0.03291822142071194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,7168,0.011719110939237805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,7168,0.19567644596099854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,5120,0.018252443936136033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,256,0.02383022175894843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,768,0.03230933348337809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,32,0.00179822225537565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,256,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,256,0.013254221942689685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,6144,0.01738755570517646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,5120,0.015992000699043274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,2560,0.009663110805882348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,32,0.001790222194459703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,5120,0.012404444317022959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,16384,0.23057956165737578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,2560,0.007340444458855524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,768,0.017685333887736004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,8192,0.012415111064910889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,1536,0.011129777464601727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,8192,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,12288,0.03216000066863166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,8192,0.029168887270821467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,12288,0.16650756200154623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,12288,0.016907556189431083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,12288,0.12522133191426596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,256,0.01512977812025282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,6144,0.02276533345381419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,256,0.004192000048028098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,5120,0.00962222201956643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,256,0.007637333538797166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,5120,0.02292266653643714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,2048,0.09560266468260024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,512,0.0027359999302360746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,2048,0.008008889026112026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,512,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,1536,0.03798666596412659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,6144,0.018252443936136033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,5120,0.008662222160233391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,12288,0.020032889313167997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,256,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,8192,0.054550220568974815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,16384,0.05805689096450806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,128,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,8192,0.011377777490350934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,16384,0.02238133384121789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,128,0.003572444534964032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,2560,0.03312089045842489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,256,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,16384,0.2471395598517524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,3584,0.0524160001013014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,1536,0.015184889237085978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,8192,0.013062222136391534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,10240,0.01238933371173011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,12288,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,65536,0.0882986651526557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,16384,0.03661422265900506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,2048,0.011328000161382886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,256,0.020101333657900494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,3584,0.007556444240940942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,128,0.010664888554149203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,512,0.012643555800120035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,7168,0.02162222233083513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,7168,0.018900444110234577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,256,0.006484444356626934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,3584,0.1035991112391154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,768,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,128,0.01705333259370592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,4096,0.015775110986497667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,8192,0.022966222630606756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,8192,0.42571022775438094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,4096,0.031102223528756037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,8192,0.010663111176755695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,10240,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,3072,0.0292195545302497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,3584,0.00998577806684706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,768,0.016233777006467182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,65536,0.0663715534740024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,10240,0.05191911260286967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,6144,0.09363910886976455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,1536,0.0034933334423436057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,768,0.015476443701320224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,3072,0.007620444728268518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,10240,0.12575556172264948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,10240,0.03551822238498264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,7168,0.19257511032952201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,12288,0.3135875595940484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,8192,0.05279199944602119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,2560,0.0058542220956749385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,16384,0.05244266655710009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,1024,0.013929777675204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,65536,1.0998062557644315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,65536,0.1841475566228231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,5120,0.02295822236273024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,2560,0.006738666858938005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,10240,0.020278222031063504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,128,0.031818668047587075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,3584,0.021173333128293354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,12288,0.04173600011401706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,5120,0.024658666716681585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,4096,0.0910408894220988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,16384,0.04383111000061035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,3072,0.08948266506195068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,128,0.007600888609886169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,65536,0.3959200117323134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,1024,0.008624888956546783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,256,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,3584,0.014152889450391134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,512,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,1024,0.0028159999185138275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,256,0.006613333192136552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,1536,0.01664177742269304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,2560,0.00700711127784517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,8192,0.01478488908873664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,4096,0.007663110891977946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,1024,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,6144,0.009648888475365108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,1024,0.00795111142926746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,256,0.002475555572244856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,8192,0.022608889473809138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,1536,0.006620444357395172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,7168,0.019887111253208585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,6144,0.07299911313586764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,1024,0.025408888856569927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,4096,0.05105422271622551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,16384,0.054661333560943604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,128,0.017511111166742113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,768,0.06753600305981107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,4096,0.008316444853941599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,64,0.0022479999396536085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,3072,0.007000000112586551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,2560,0.08584800031450059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,12288,0.05810222360822889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,1024,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,3584,0.008603555460770925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,8192,0.01808888879087236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,2048,0.055481778250800244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,10240,0.04157600137922499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,6144,0.03354133168856303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,1024,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,2048,0.006559999866618051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,1536,0.017879999346203275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,1536,0.015832000308566622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,4096,0.004607111215591431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,512,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,768,0.01071111112833023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,4096,0.0236817780468199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,2560,0.007280000381999546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,10240,0.03176800078815884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,256,0.005623111294375525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,1024,0.06286844280030993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,512,0.010020444790522257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,512,0.013099555340078143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,8192,0.029178665743933782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,10240,0.06419822242524889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,64,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,32,0.005642666584915585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,4096,0.003914666672547658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,16384,0.02748799986309475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,512,0.005495999836259418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,10240,0.013755555782053204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,4096,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,256,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,10240,0.5285040007697211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,2048,0.004511111312442356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,16384,0.03230133321550157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,256,0.011563555234008364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,4096,0.023020444644822016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,128,0.013454222016864352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,4096,0.06550488869349162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,256,0.015127110812399121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,1536,0.01274222218328052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,10240,0.24377332793341744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,32,0.006604444649484422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,3584,0.007968000239796108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,128,0.002489777695801523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,1536,0.006984000404675801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,1536,0.012406222522258759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,2560,0.012443555725945367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,512,0.0028355556229750314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,2560,0.025656888882319134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,10240,0.020675554871559143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,512,0.015954666667514376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,512,0.0026088888860411118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,10240,0.004233777936961916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,5120,0.011580444044537015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,10240,1.118600845336914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,65536,0.47764354281955296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,5120,0.011776888536082374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,10240,0.3725431230333116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,12288,0.01516266663869222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,1024,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,2048,0.01644533375898997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,512,0.01111822244193819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,3584,0.01349422252840466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,16384,0.013559111290507846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,768,0.006205333189831839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,4096,0.0733484427134196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,12288,0.005532444351249271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,12288,0.013627555635240344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,64,0.0031431110368834603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,512,0.005920889063013925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,10240,0.03858400053448147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,1536,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,6144,0.010119111173682744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,3072,0.10681777530246311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,4096,0.012715555727481842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,10240,0.03124799993303087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,1024,0.015119110544522604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,128,0.012860444684823355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,5120,0.02680088910791609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,8192,0.05836621920267741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,32,0.0053084443012873335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,768,0.005984888722499211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,2048,0.01758311192194621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,1024,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,5120,0.024688000480333965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,2560,0.08975111113654242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,2048,0.0123831108212471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,1024,0.007084444165229797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,7168,0.03289866778585646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,16384,0.14704622162712946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,3072,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,32,0.002513777878549364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,16384,0.051552888419893056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,7168,0.018530666828155518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,2048,0.014107555150985718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,6144,0.0092275556590822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,3584,0.008348444269763099
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,5120,0.011329777538776398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,6144,0.023727110690540735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,12288,0.04507466819551256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,8192,0.013723555538389417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,5120,0.009018667042255402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,2560,0.004910222358173794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,8192,0.023784889115227595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,3072,0.005519999812046687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,512,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,768,0.009029332962301042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,16384,0.13417422771453857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,8192,0.045924444993337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,1536,0.07872977521684435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,16384,0.04584088921546936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,1024,0.014801777071423002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,6144,0.006312888943486744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,256,0.016819554898473952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,6144,0.009722666607962714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,768,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,8192,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,65536,0.0973893337779575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,8192,0.03254844413863288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,1024,0.008263111114501953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,10240,0.028619554307725694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,128,0.00217333332531982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,3584,0.010605333579911126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,32,0.011764444410800934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,256,0.006664888726340399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,64,0.0018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,8192,0.01959022217326694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,2560,0.012576888832781049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,3072,0.007443555527263218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,3072,0.017666666044129264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,8192,0.011015999648306104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,2560,0.007532444265153673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,3072,0.02330933345688714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,2560,0.03515200151337518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,2560,0.25605776574876576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,2048,0.008101333346631791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,3584,0.1879333390129937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,128,0.002529777793420686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,12288,0.06095200114780002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,256,0.04351466563012865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,512,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,12288,0.0396088891559177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,5120,0.017864889568752713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,1024,0.006190222170617845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,4096,0.18498044543796113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,5120,0.05770133601294624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,2560,0.017050666941536795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,16384,0.06674933433532715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,2048,0.012375999655988483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,64,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,1536,0.014122666584120857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,6144,0.019295111298561096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,64,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,256,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,65536,0.050256000624762646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,1536,0.024500444531440735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,6144,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,3072,0.02862488892343309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,1536,0.10758666859732734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,12288,0.05349599983957079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,8192,0.023690667417314317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,6144,0.0273973329199685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,3072,0.025359110699759588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,2560,0.007333333293596904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,5120,0.03508799937036302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,8192,0.07160622543758817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,16384,1.1693955527411568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,6144,0.03904977771970961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,2048,0.008950221869680617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,32,0.007663110891977946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,64,0.008005333443482717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,64,0.003522666792074839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,6144,0.018534221582942538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,3072,0.028375999795065984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,64,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,8192,0.029923554923799302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,4096,0.09521244631873237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,8192,0.01570311188697815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,12288,0.016925333274735343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,12288,0.06222666634453667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,10240,0.021535999245113794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,3584,0.09757866462071736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,256,0.014803555276658801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,4096,0.06317777766121759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,6144,0.02156977852185567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,1536,0.01681422193845113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,7168,0.013312000367376538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,65536,0.07575911283493042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,3072,0.009054222040706212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,768,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,1536,0.015492444237073263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,512,0.0028355556229750314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,7168,0.028762665059831407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,256,0.018011555075645447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,768,0.002883555574549569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,3584,0.019482667247454327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,768,0.0057839999596277876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,2048,0.07660977707968818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,65536,0.08264355527030097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,6144,0.024666666984558105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,1536,0.005909333212508096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,10240,0.017027555240525138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,16384,0.01814311080508762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,3072,0.007761778102980719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,12288,0.027335110637876723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,6144,0.004888888862397936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,3584,0.007692444655630324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,6144,0.14429244730207655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,4096,0.0041884444653987885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,12288,0.06718844175338745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,1536,0.006618666566080517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,3072,0.012313777373896705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,8192,0.015493333339691162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,3584,0.015146666102939181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,768,0.003524444583389494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,65536,0.12515111764272055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,16384,0.03635288940535651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,2560,0.006979555719428592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,512,0.010563555690977307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,256,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,3584,0.019909333851602342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,2560,0.03398222062322829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,256,0.0024951110697454875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,64,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,128,0.0053377776510185665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,768,0.014787554740905762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,16384,0.01478844384352366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,128,0.005828444328573015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,512,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,5120,0.009837333526876237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,1024,0.016954667038387723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,2560,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,32,0.00249777775671747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,12288,0.10793333583407932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,32,0.010324444207880232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,768,0.005613333235184352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,10240,0.1749546660317315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,768,0.01664622293578254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,32,0.010332444475756751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,6144,0.38632978333367246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,1536,0.013793778088357715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,4096,0.22160177760654023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,1024,0.019002666076024372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,2560,0.007338666253619724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,7168,0.015298666225539314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,4096,0.028755555550257366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,12288,0.032594667540656194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,3072,0.010320000350475311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,64,0.002303999952144093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,7168,0.013615111509958902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,2048,0.05198755529191759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,8192,0.032483554548687406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,256,0.009688888986905416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,3072,0.010348444183667501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,256,0.022058667408095464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,3584,0.017846221725145977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,128,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,256,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,512,0.002492444382773505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,512,0.011711999773979187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,512,0.005636444522274866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,65536,0.0872924460305108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,6144,0.024855999483002558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,128,0.0038844446341196695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,1024,0.022687110635969374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,768,0.0032222221295038858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,7168,0.013060444758998023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,10240,0.1412444379594591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,2560,0.01167022188504537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,256,0.005607111172543631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,4096,0.012441777520709567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,2560,0.014091556270917257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,768,0.007255111303594377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,64,0.01755555636352963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,10240,0.012100444071822695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,128,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,2560,0.017084444562594097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,2560,0.007615110940403408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,2560,0.00702400008837382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,2560,0.013803555733627744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,4096,0.026397332549095154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,768,0.022982221510675218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,1536,0.034852445125579834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,5120,0.05531555414199829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,10240,0.039080000585979886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,2560,0.007621333003044128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,768,0.006250666661394968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,512,0.007610666255156199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,6144,0.013034666577974955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,10240,0.03679022192955017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,7168,0.021597334080272253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,8192,0.018868444694413077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,16384,0.032820443312327065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,2560,0.020241777102152508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,512,0.002487999904486868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,10240,0.6692399978637695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,4096,0.026847110854254827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,512,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,5120,0.006562666760550604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,64,0.0018008889423476325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,256,0.015842666228612263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,1024,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,5120,0.008577778107590145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,3072,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,10240,0.010581333604123859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,1024,0.007306666837798224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,3072,0.005028444445795483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,65536,0.5548382335238987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,128,0.015463999576038785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,12288,0.022244445151752893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,5120,0.012423111332787408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,3072,0.010078222387366826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,16384,0.015835555063353646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,256,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,12288,0.21191378434499106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,65536,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,4096,0.016857778032620747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,12288,0.013413333230548434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,7168,0.021928000781271193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,512,0.010713777608341642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,64,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,1024,0.006119110931952794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,10240,0.01202400028705597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,16384,0.015844444433848064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,2560,0.06105066670311821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,128,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,3072,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,256,0.0021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,12288,0.07810133033328585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,4096,0.004586666822433472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,512,0.016713778177897137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,1536,0.008647110727098253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,3072,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,3584,0.007649777664078607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,1024,0.0035537779331207275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,12288,0.03015111221207513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,3584,0.009901333186361525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,256,0.00519822206762102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,6144,0.015005333556069268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,128,0.015113777584499784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,3072,0.007443555527263218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,4096,0.012719111310111152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,8192,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,12288,0.00452622233165635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,1024,0.0236488895283805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,5120,0.009436444275909001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,2560,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,256,0.008266666697131263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,3072,0.020224000016848247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,7168,0.04381688767009311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,128,0.005863111052248213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,6144,0.013415111435784234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,6144,0.13884266217549643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,3584,0.007460444337791867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,1536,0.007748444047239091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,128,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,12288,0.042452444632848106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,1024,0.013059555656380124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,512,0.03242133392228021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,3072,0.013730666703648038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,256,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,1024,0.012247111234400006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,65536,0.05962044662899441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,10240,0.033872889147864446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,3072,0.041119111908806696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,3584,0.01648355523745219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,12288,0.027069333526823256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,2048,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,4096,0.0039022221333450745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,512,0.03420799970626831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,4096,0.008226667013433244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,3072,0.09523200326495701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,2560,0.004903111192915174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,1536,0.007918222082985772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,256,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,7168,0.011717333561844297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,8192,0.08249866962432861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,3072,0.010677333507272931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,256,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,1536,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,128,0.0023440000497632557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,7168,0.014681778020328946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,4096,0.00452533322903845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,5120,0.00489155575633049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,7168,0.016318221886952717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,7168,0.06470133198632134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,6144,0.08366400003433228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,32,0.01924622224436866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,65536,0.3537031014760335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,16384,0.060384001996782094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,32,0.004929777648713854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,6144,0.0553653339544932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,1536,0.014488000008794995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,1536,0.006284444282452266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,16384,0.022654222117529974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,7168,0.05428266525268555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,2560,0.0075839997993575195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,3584,0.02681688964366913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,10240,0.03894133369127909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,2048,0.007703111403518253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,256,0.015501333607567681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,4096,0.00868444475862715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,128,0.0021280000607172647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,65536,0.08353866471184625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,2560,0.005502222312821283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,12288,0.00902133352226681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,3072,0.16974755128224692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,768,0.0029760001020299066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,4096,0.008319999608728621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,768,0.025798221429189045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,3072,0.009679999616410997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,5120,0.008972444468074376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,2048,0.06643644306394789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,12288,0.033230221933788724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,2048,0.042223999897638954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,128,0.0147706667582194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,3584,0.17474044693840873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,4096,0.022688888841205176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,8192,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,10240,0.016935111747847665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,2048,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,2048,0.017512889371977914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,2560,0.004908444566859139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,4096,0.01811733345190684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,3584,0.008193777667151557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,8192,0.025057777762413025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,5120,0.12035022841559516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,2048,0.006904888898134232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,5120,0.012758221891191272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,2048,0.019934222102165222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,10240,0.48665957980685765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,4096,0.19467111428578696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,2560,0.03085422184732225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,128,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,1536,0.015824889143308003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,8192,0.7272657818264432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,7168,0.013041777743233575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,4096,0.12407910823822021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,16384,0.033385776811175875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,1024,0.024005333582560223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,7168,0.03695288962788052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,7168,0.010316444767846001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,7168,0.1579173273510403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,3584,0.10187733173370361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,3072,0.02198133369286855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,512,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,768,0.017840888765123155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,1536,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,8192,0.3826773431566026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,16384,0.2075208822886149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,3584,0.014766222900814481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,6144,0.009899555808968013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,128,0.005495111147562663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,256,0.005103111267089844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,768,0.0027991111079851785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,7168,0.011385777758227454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,65536,0.08666666348775227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,7168,0.010724444356229572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,2048,0.00796088907453749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,512,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,12288,0.0373288889726003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,256,0.01513155632548862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,128,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,768,0.005423999908897612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,10240,0.22772799597846138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,768,0.004871110949251386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,2560,0.004555555681387584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,6144,0.011004444625642566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,12288,0.03055822186999851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,3584,0.008709333009190029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,6144,0.06811466481950548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,16384,0.6394133567810059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,8192,0.04182933436499702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,2560,0.02914311157332526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,5120,0.02128533356719547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,768,0.011053333679835001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,8192,1.5349146525065105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,32,0.008957333034939235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,65536,0.017335999343130324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,2048,0.022671999202834234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,256,0.007391999993059371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,2560,0.011063999599880643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,1024,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,10240,0.012979555461141797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,6144,0.007420444654093848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,512,0.0075297777851422625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,10240,0.02144977781507704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,2048,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,1536,0.010330666270520952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,1024,0.027002667387326557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,16384,0.0547288888030582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,2048,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,768,0.040425777435302734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,1536,0.0068862222962909276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,3584,0.013029333617952136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,3072,0.010794666906197866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,768,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,10240,0.013437333206335703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,256,0.011707555916574268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,3584,0.015522667103343539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,3584,0.014212444424629211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,768,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,1024,0.011715555356608497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,65536,0.07161955700980292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,5120,0.008350222474998897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,5120,0.010806221928861407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,256,0.002524444419476721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,2048,0.004549333204825719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,12288,0.018090666996108163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,12288,0.01755911111831665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,1024,0.005927111125654644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,768,0.0028808888875775864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,65536,0.02555199960867564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,2048,0.0594959987534417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,65536,0.05768711037105984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,3584,0.027064889669418335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,8192,0.02001688877741496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,65536,0.2582968870798747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,768,0.006893333461549547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,65536,0.10450844632254706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,32,0.0017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,4096,0.030187557140986126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,7168,0.012747555143303342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,2048,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,6144,0.008964444200197855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,1536,0.0062871111763848206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,2560,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,8192,0.18155999978383383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,2560,0.04168355464935303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,12288,0.013394667042626275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,12288,0.04437866806983948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,2560,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,8192,0.023315555519527856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,128,0.02015022271209293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,1024,0.007596444752481248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,2048,0.004544888933499654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,512,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,2048,0.016862221890025668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,32,0.0021404443929592767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,1536,0.01752000053723653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,2560,0.013591110706329346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,3072,0.030044443077511255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,1024,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,3072,0.009701333112186855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,2560,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,3584,0.011039999624093374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,1536,0.003240888938307762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,65536,0.061082667774624295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,128,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,256,0.004947555561860402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,7168,0.04485422372817993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,1024,0.008294222255547842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,2048,0.006650666809744305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,4096,0.007651555869314406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,3072,0.00847999999920527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,768,0.005187555733654234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,2048,0.004217777815130022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,128,0.011695110963450538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,65536,0.09498933288786147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,16384,0.09378488858540852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,256,0.027179555760489568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,512,0.002620444529586368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,256,0.006280888699822956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,10240,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,12288,0.028060442871517603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,10240,0.03277600142690871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,5120,0.008621333373917473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,4096,0.061828447712792285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,12288,0.034663110971450806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,16384,0.30178578694661456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,7168,0.04157688882615831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,1024,0.043486220969094165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,10240,0.35374755329555935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,3584,0.04231199953291151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,10240,0.015862221519152325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,1536,0.010673777924643623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,128,0.040001776483323835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,65536,0.1345911158455743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,8192,0.04874311221970452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,5120,0.008991111483838823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,3584,0.03760000069936117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,128,0.013507555756303998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,64,0.0031519999934567344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,256,0.006575111299753189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,6144,0.034274667501449585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,7168,0.15589777628580728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,512,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,7168,0.011360000405046674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,256,0.007228444847795699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,1536,0.024711999628278945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,256,0.0024213333510690266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,8192,0.041088001595603094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,2560,0.006959111326270633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,12288,0.050372444921069674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,3584,0.015016888578732809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,1536,0.018897778458065458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,2560,0.13891200224558511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,3584,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,1536,0.00889066689544254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,1024,0.0028444443725877335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,256,0.002136888810329967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,3584,0.013763555222087435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,64,0.008266666697131263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,2048,0.10920000076293945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,3072,0.09806399875217015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,6144,0.010136889086829292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,512,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,5120,0.010020444790522257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,2560,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,12288,0.25388532214694554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,2048,0.012031111452314587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,65536,0.12251110871632893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,128,0.0028515555378463534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,4096,0.014496889379289416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,12288,0.3005724483066135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,3072,0.11525244183010525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,256,0.007794666621420119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,1024,0.013177778157922955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,512,0.008635555704434713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,2048,0.008087111016114553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,12288,0.025064888927671645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,3072,0.02257866660753886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,64,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,4096,0.007972444097201029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,1536,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,65536,0.12415822347005208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,1536,0.06879378027386136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,2560,0.009670221971140968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,2560,0.017281777328915067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,32,0.0052986666560173035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,64,0.011220444407727985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,5120,0.009650666680600908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,512,0.006980444822046492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,4096,0.021624000536070928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,8192,0.016531555189026725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,3584,0.008832000195980072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,4096,0.02344266739156511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,6144,0.004949333353175057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,5120,0.013757333159446716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,7168,0.023778667052586872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,2560,0.006924444602595435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,3584,0.03476444549030728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,768,0.020301333732075162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,1024,0.015504888362354703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,12288,0.017613333132531907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,2560,0.02423288921515147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,12288,0.215192887518141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,256,0.0026977778308921387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,4096,0.016180444094869826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,10240,0.06389066908094618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,3072,0.03557599915398492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,12288,0.02074666652414534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,7168,0.017872888180944655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,1536,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,2560,0.007215110792054071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,3072,0.015509333875444202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,10240,0.20931643909878203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,512,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,10240,0.033999999364217125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,128,0.01648000048266517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,4096,0.01424888935354021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,2048,0.02772177755832672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,1536,0.020404444800482858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,16384,0.4147786564297146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,3072,0.12921599547068277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,3072,0.008744888835483128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,8192,0.011353777514563667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,2048,0.011072888970375061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,6144,0.5324959754943848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,65536,0.07614311244752672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,768,0.009367110828558603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,6144,0.028393778536054824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,768,0.017204445269372728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,8192,0.03977955712212457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,32,0.006414222220579783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,3072,0.008993777963850234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,5120,0.029162665208180744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,65536,0.10436622301737468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,8192,0.013582222163677216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,2048,1.5002479553222656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,65536,0.08208977513843112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,2048,0.018757333358128864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,6144,0.06842133071687487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,10240,0.03646399908595615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,2048,0.015160888433456421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,512,0.003157333367400699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,1024,0.0747786694102817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,65536,0.21963911586337617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,5120,0.017113778326246474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,2048,0.004563555535342959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,2048,0.014448000325096978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,1024,0.01295377810796102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,768,0.004545777622196409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,7168,0.020595555504163105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,32,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,2560,0.026408889227443274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,1024,0.016248888439602323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,65536,0.11682311693827312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,1536,0.006288888967699475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,6144,0.02659911082850562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,10240,0.05847822295294868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,12288,0.010616888602574667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,12288,0.04647289050949944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,65536,0.10375733508004083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,7168,0.016021332807011075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,32,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,8192,0.04478933413823446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,2560,0.012995555169052549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,12288,0.018178666631380718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,64,0.00198755558166239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,3072,0.5146302117241753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,4096,0.016509332590632968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,2560,0.004907555464241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,6144,0.005974222388532426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,768,0.0032026666320032547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,1536,0.023333332604832117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,1024,0.006269333263238271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,65536,0.05660088857014974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,2560,0.02110133402877384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,1536,0.015473778049151102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,2560,0.013473778135246701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,3584,0.02605155607064565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,5120,0.00832088871134652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,512,0.005292444593376583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,12288,0.013821333646774292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,512,0.0041982221106688184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,4096,0.022625777456495497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,128,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,64,0.013337777720557319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,16384,0.041033777925703264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,64,0.001842666624320878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,6144,0.009390222529570261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,10240,0.01388533330625958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,8192,0.016354666815863717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,5120,0.007968000239796108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,3584,0.011852444046073489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,4096,0.021572444174024794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,4096,0.011709333293967776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,2560,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,1536,0.004231111043029361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,512,0.00250133333934678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,768,0.027815110153622095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,4096,0.047961778110928006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,256,0.005961777849329843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,6144,0.012772444221708508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,64,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,6144,0.034620443979899086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,256,0.016176889340082806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,5120,0.008620444271299574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,3584,0.024000888069470722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,6144,0.010412444670995077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,65536,2.12028439839681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,512,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,10240,0.03281689021322463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,32,0.013080889152155982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,4096,0.044863111442989774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,16384,0.020576889316240948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,32,0.0019253332995706135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,7168,0.011730666789743634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,768,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,16384,0.047037333250045776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,64,0.0021555556191338436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,65536,0.08637422323226929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,2048,0.012944888737466602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,12288,0.02903644575013055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,16384,0.052707556221220225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,2560,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,128,0.0151555554734336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,3584,0.3284817801581489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,2560,0.022350221872329712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,6144,0.12042310502794053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,1024,0.007344889144102733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,768,0.011042666931947073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,2048,0.06051999992794461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,16384,0.07818488942252265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,512,0.025479111406538222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,1536,0.003310222178697586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,8192,0.16322755813598633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,2048,0.00793511089351442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,768,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,6144,0.03161777721510993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,7168,0.03852444556024339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,16384,0.038600001070234514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,12288,0.18592800034417045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,3584,0.015450666348139444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,4096,0.02046133412255181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,3584,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,7168,0.006999111009968652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,2048,0.024028443627887305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,1024,0.023631110787391663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,1024,0.004062222109900581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,64,0.006411555740568373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,5120,0.028879112667507593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,3072,0.009683555199040307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,5120,0.03094489044613308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,1536,0.006055111272467508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,256,0.014864888456132678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,3584,0.007658667034573025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,3072,0.007802666889296637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,2048,0.012420444852775998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,128,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,256,0.009368889033794403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,2048,0.01648622254530589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,7168,0.011244444383515252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,32,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,5120,0.010035555395815108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,128,0.006309333360857434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,4096,0.20306132899390328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,256,0.012060444388124677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,16384,0.04833333359824287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,10240,0.026359111070632935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,65536,0.17061777909596762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,128,0.0021724444296624926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,3584,0.011029333704047732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,64,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,128,0.015449777245521545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,1536,0.012049777640236748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,12288,0.1483022239473131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,6144,0.015777778294351365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,7168,0.011595555477672152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,10240,0.24968443976508248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,6144,0.00999644481473499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,4096,0.007992888490358988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,3584,0.01593955523437924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,16384,0.03577955563863119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,1536,0.012741333080662621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,2560,0.014696889453464083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,16384,0.019941333267423842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,256,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,512,0.00574222207069397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,128,0.1069342229101393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,256,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,512,0.013044444223244986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,2560,0.015483554866578845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,5120,0.013065777719020844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,1024,0.002887111157178879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,512,0.002507555608948072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,256,0.006639110959238476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,8192,0.018252443936136033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,768,0.010097777677906884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,4096,0.015112888481881885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,10240,0.031039112144046362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,3072,0.020959110723601446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,3584,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,64,0.05438666542371114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,8192,0.04080444574356079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,6144,0.018001778258217704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,7168,0.01629155543115404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,10240,0.020693333612547982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,16384,0.08840888738632202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,32,0.012715555727481842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,7168,0.24715733528137207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,6144,0.1178648869196574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,512,0.005952888892756567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,32,0.007612444460391998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,5120,0.06278044647640653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,16384,0.03811555438571506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,8192,0.04749511016739739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,2560,0.007319110963079665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,16384,0.264627562628852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,3584,0.018648000227080453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,65536,0.03756444321738349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,2560,0.0070257774657673305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,3072,0.04823199907938639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,6144,0.004879111217127906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,2560,0.007240888973077138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,1024,0.006095110956165526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,3584,0.004898666507667965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,1024,0.00630666646692488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,768,0.017161778277821012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,12288,0.014800889624489678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,65536,0.06382844183180067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,12288,0.04509422183036804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,768,0.015850666496488783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,6144,0.2106693320804172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,2560,0.007367110914654202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,256,0.01647733317481147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,512,0.01423377792040507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,2048,0.0620408919122484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,1024,0.02308533257908291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,512,0.022428444690174524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,1024,0.04754755563206143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,32,0.01683288812637329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,64,0.001839111041691568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,8192,0.11520089043511285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,1536,0.0035297779573334586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,512,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,12288,0.016894222961531747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,256,0.0021626665774318907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,3584,0.011371555427710215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,3584,0.01869244376818339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,5120,0.12425866391923691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,4096,0.007967111137178209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,768,0.002507555608948072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,128,0.0032008888406885993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,5120,0.006613333192136552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,512,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,8192,0.010686222049925061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,16384,0.011499555574523078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,1024,0.0032248888164758682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,3072,0.0069155556460221606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,5120,0.033547557062572904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,5120,0.009005332986513773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,64,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,1024,0.006625777731339137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,65536,0.11466311083899604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,5120,0.06717511018117268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,3584,0.007910221815109253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,12288,0.01649688846535153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,3584,0.02258844508065118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,1024,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,128,0.005968889014588461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,6144,0.017494221528371174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,6144,0.03703377644220988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,64,0.0018097776919603348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,6144,0.01036622209681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,512,0.013423110875818463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,2560,0.059000889460245766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,2048,0.006934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,4096,0.01389955480893453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,1536,0.005527110977305307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,3584,0.00795199970404307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,16384,0.02615822189384037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,65536,0.06447822517818876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,32,0.008627555436558193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,65536,0.1461306677924262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,512,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,4096,0.006088888893524806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,8192,0.07041333119074504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,256,0.0766213337580363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,6144,0.010008888940016428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,768,0.01290844463639789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,7168,0.015458666616015963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,16384,0.04345422320895725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,1024,0.003530666646030214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,8192,0.010878221856223213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,4096,0.014836443795098199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,64,0.04425599839952257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,8192,0.3884595500098334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,3072,0.015271110667122735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,2560,0.07899200254016452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,1536,0.12487910853491889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,7168,0.02478133307562934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,2560,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,1024,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,768,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,768,0.013041777743233575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,65536,0.16655645105573866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,256,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,6144,0.007631999750932057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,2048,0.004860444615284602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,2048,0.017200888858901132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,10240,0.012801777985360889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,768,0.005038222091065513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,2048,0.08105422390831842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,65536,0.1833724445766873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,4096,0.01033688916100396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,3584,0.007885333564546373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,16384,0.020562666985723708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,8192,0.04342933495839437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,10240,0.10811644130282932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,12288,0.05192266570197212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,7168,0.03436977664629618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,65536,0.06394933329688178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,128,0.015856888559129503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,64,0.3243617746565077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,2048,0.005956444475385878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,7168,0.1939368910259671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,128,0.005920889063013925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,3584,0.010999110837777456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,256,0.012095111111799875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,256,0.00534400012758043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,256,0.008273777862389883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,3072,0.012554667062229581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,3584,0.01071200023094813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,4096,0.04927466644181145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,2560,0.014919999572965833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,128,0.00552977787123786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,6144,0.01038044442733129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,768,0.026664889521068994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,65536,0.16195555528004965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,256,0.017383111847771537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,768,0.017449777987268236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,3584,0.017771555317772757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,10240,0.014095111025704278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,1536,0.02051555613676707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,512,0.012082666986518435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,10240,0.015804444750150044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,1536,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,3584,0.007947555846638149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,16384,0.05993511279424032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,512,0.004902222090297275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,5120,0.010001777774757808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,512,0.006018666757477655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,2560,0.007651555869314406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,5120,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,512,0.01347644461525811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,128,0.011777777638700275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,12288,0.011718221836619906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,8192,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,1024,0.005757333503829108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,3584,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,512,0.015531554818153381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,12288,0.02986933456526862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,1024,0.040467556979921125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,12288,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,32,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,64,0.0025351111673646504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,3072,0.006224000205596288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,8192,0.011578666667143503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,4096,0.021333333518770006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,1024,0.0032302221904198327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,6144,0.009650666680600908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,2560,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,512,0.013449778159459433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,12288,0.025507556067572698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,1536,0.012958221965365939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,2560,0.13895644081963435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,1536,0.015420445137553744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,12288,0.03145600027508206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,5120,0.006566222343179915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,3584,0.051346666283077665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,65536,6.164777967664931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,10240,0.10902933279673259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,8192,0.017857778403494094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,65536,0.1447333362367418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,128,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,1536,0.02879377868440416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,8192,0.02241955531968011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,1536,0.01716444392999013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,6144,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,7168,0.009693333672152625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,1536,0.015174221661355762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,65536,0.10903733306460911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,5120,0.0053119998839166425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,65536,0.05438933438724942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,128,0.005218666460778978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,3072,0.015428443749745687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,3072,0.00424177779091729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,65536,0.0795102251900567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,6144,0.02971822354528639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,5120,0.009816889133718278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,5120,0.010009778042634329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,1536,0.0047182221379545005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,3072,0.004603555632962121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,2560,0.009706666900051964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,3072,0.008264888491895463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,16384,0.09054044220182632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,256,0.014837332897716098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,512,0.019229332605997723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,8192,0.023396443989541795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,4096,0.01832177738348643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,65536,0.22161955303615996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,10240,0.048067556487189404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,1024,0.012496000362767113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,1536,0.02478844424088796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,5120,0.008198222352398766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,32,0.0031822222388452957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,2048,0.023784000012609694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,7168,0.1399164464738634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,512,0.005752888818581899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,2560,0.0629262195693122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,512,0.006253333141406377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,4096,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,1536,0.07164355781343248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,32,0.007978666987684038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,5120,0.0477182235982683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,512,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,256,0.002480888946188821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,8192,0.020236444142129686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,64,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,1536,0.006503110958470239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,2560,0.017157332764731515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,4096,0.016659556163681876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,64,0.0036720000207424164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,256,0.005385777602593104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,3584,0.019550222489568923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,512,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,8192,0.03247911069128249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,1536,0.01978222197956509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,2048,0.020192000601026747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,65536,0.15025688542260066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,256,0.0022026666750510535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,16384,0.05041422115431892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,3072,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,3584,0.007927999728255803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,2048,0.012660444610648684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,32,0.0041893331540955436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,65536,0.06491466363271077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,128,0.0021262222694026097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,2048,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,512,0.0028817777832349143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,768,0.0028328889360030494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,64,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,64,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,10240,1.8499253590901692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,1536,0.009992889232105678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,16384,0.45246219635009766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,12288,0.22476000256008574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,1024,0.0069875551594628235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,16384,0.4665502442253961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,16384,0.060976001951429576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,8192,0.011074666347768573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,64,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,6144,0.020336889558368258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,1536,0.007131555842028723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,8192,0.01921244462331136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,4096,0.007961778177155389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,1536,0.0038328886859946777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,2048,0.009387555221716562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,64,0.002191999927163124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,2048,0.01715288890732659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,512,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,3072,0.013655111193656921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,2048,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,128,0.0021564445147911706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,3584,0.008824889030721452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,64,0.002032888846264945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,64,0.007886221839321984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,3072,0.039319998688168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,4096,0.04350489046838549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,16384,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,7168,0.009127111070685917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,65536,0.05685422155592176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,10240,0.05264177918434143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,8192,0.028456889920764502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,256,0.016858667135238647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,128,0.003481777591837777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,1024,0.0028844444702068963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,4096,0.008628444539176093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,3584,0.019119999474949308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,1536,0.019707555572191875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,6144,0.004924444274769889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,8192,0.013182222015327878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,10240,0.10735911130905151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,1536,0.014485332700941296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,2048,0.02068088948726654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,128,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,12288,0.004866666677925321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,768,0.007278222176763747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,2560,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,3072,0.02882933285501268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,256,0.002209777840309673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,8192,0.043920000394185386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,3072,0.01372977760103014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,7168,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,128,0.005222222043408289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,16384,0.03717689050568475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,256,0.0028826666788922418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,16384,0.06575466526879205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,2048,0.004243555582231945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,4096,0.033415112230512835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,3072,0.08498933580186631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,3072,0.0123848890264829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,7168,0.02176622218555874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,128,0.005208000126812193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,8192,0.02090488870938619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,6144,0.009650666680600908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,4096,0.006618666566080517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,6144,0.022851554883850947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,10240,0.015749333633316886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,512,0.012065778175989786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,5120,0.016144000821643405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,6144,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,64,0.0022444443570242985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,8192,0.22036533885531953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,10240,0.051270223326153226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,1024,0.004498666773239772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,2560,0.019612444771660697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,10240,0.014875556031862894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,3072,0.007282666862010956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,64,0.005944889038801193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,6144,0.020894222789340552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,5120,0.012326222327020435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,12288,0.06991377804014418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,1024,0.024659555819299486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,8192,0.04727733466360304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,5120,0.03918577896224128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,7168,0.020814221766259935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,65536,0.20982666810353598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,3072,0.028573334217071533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,16384,0.09018755621380276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,64,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,512,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,2560,0.006956444846259222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,768,0.016193777322769165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,8192,0.03224444389343262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,64,0.0018844445132546956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,6144,0.052630222505993314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,4096,0.022038223014937505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,8192,0.6571120156182183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,16384,0.0756959981388516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,128,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,6144,0.02390488816632165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,65536,0.34006399578518337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,256,0.01107377807299296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,16384,0.0057911111248864075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,512,0.014018666413095264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,2048,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,8192,0.17318667305840385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,10240,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,3584,0.008642666869693333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,512,0.016179554992251925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,7168,0.020436444216304354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,4096,0.010680889089902243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,16384,0.047621332936816745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,3584,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,12288,0.013743110828929476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,2048,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,5120,0.009043555292818282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,7168,0.012256888879670037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,768,0.12572533554501003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,4096,0.015413333972295126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,16384,0.35866488350762266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,12288,0.18415733178456625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,3584,0.007583111524581909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,5120,0.00833777752187517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,8192,0.022673777408070032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,6144,0.027093332674768236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,768,0.009354666703277165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,8192,0.011048888994587792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,10240,0.02254933284388648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,128,0.01332800007528729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,512,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,12288,0.016583111551072862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,3584,0.1024435559908549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,10240,0.02574399941497379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,8192,0.010654222634103564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,12288,0.016583111551072862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,3072,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,512,0.009845332966910468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,8192,0.0777182247903612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,5120,0.010075555907355415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,512,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,16384,0.019707555572191875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,768,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,16384,9.63222927517361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,2560,0.02535199953450097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,64,0.009680888719028896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,64,0.008965333302815756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,512,0.021609778205553692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,1024,0.031034668286641438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,8192,0.02197511163022783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,1024,0.006903999795516332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,6144,0.0159582214223014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,65536,0.2284888956281874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,1024,0.015117333994971381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,512,0.01478488908873664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,64,0.014105778601434497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,10240,0.027975112199783325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,12288,0.018790221876568265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,3072,0.013148444394270578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,256,0.014075555735164218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,10240,0.016895999511082966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,12288,0.014072888427310519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,1536,0.006530666516886816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,10240,0.015784889459609985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,16384,0.2616186671786838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,12288,0.04530577858289083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,2048,0.010100444157918295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,64,0.00179822225537565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,768,0.01718844473361969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,768,0.009644444617960189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,65536,0.04688533478313022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,2048,0.01258044441541036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,12288,0.6904213163587781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,7168,0.012717333104875354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,4096,0.13569866286383733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,3072,0.18770578172471789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,512,0.04037422272894118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,3072,0.01383733335468504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,768,0.01757688820362091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,512,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,2560,0.027929776244693335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,64,0.008994667066468133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,1024,0.009326222042242685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,6144,0.011309333145618439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,1536,0.01309511148267322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,1536,0.006685333533419504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,3584,0.015503999259736804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,65536,0.06334488921695285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,5120,0.0045422220395671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,5120,0.017217778497272067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,3584,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,768,0.00612088872326745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,256,0.013393777940008374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,1536,0.009352888498041365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,256,0.01681600014368693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,8192,0.011043555206722684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,3072,0.011126221881972419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,12288,0.035339554150899254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,1024,0.0148026661740409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,32,0.002335999988847309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,7168,0.019241778386963736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,3584,0.07206755876541138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,16384,0.049913777245415576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,2560,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,128,0.015457777513398064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,64,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,16384,0.021798221601380244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,2560,0.007596444752481248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,8192,0.13448088698916966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,1024,0.0255377772781584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,3584,0.012006222373909421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,256,0.029762665430704754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,7168,0.061088886525895864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,3072,0.054494222005208336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,3072,0.09022844500011867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,5120,0.01758666667673323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,2560,0.14815733167860243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,768,0.00573955559068256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,3072,0.01443733274936676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,7168,0.018167111608717177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,3072,0.01646844380431705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,8192,0.017109332813156974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,512,0.005251555393139522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,10240,0.011040888726711273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,512,0.01891911029815674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,4096,0.03103999959097968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,10240,0.0151555554734336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,64,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,7168,0.01071733319097095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,3584,0.02920888861020406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,2048,0.020392000675201416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,512,0.0035342222286595237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,2560,0.010663111176755695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,10240,0.023346667488416035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,65536,0.053322666221194796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,1536,0.007807999849319458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,12288,0.13746933142344156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,12288,0.026312000221676294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,4096,0.02425333360830943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,12288,0.17419021659427217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,6144,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,6144,0.020050666398472257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,128,0.004912888838185204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,1024,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,2048,0.021406221720907424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,10240,0.011717333561844297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,6144,0.009352000223265754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,7168,0.013540444274743399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,512,0.06660355461968316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,256,0.015842666228612263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,128,0.0018444444156355327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,512,0.022451554735501606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,512,0.0021928888228204516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,65536,0.20344177881876627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,65536,0.16203555795881483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,768,0.012840888566441007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,2560,0.007481777833567725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,32,0.002522666628162066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,5120,0.018608000543382432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,16384,0.00599911105301645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,3584,0.004578666554556953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,256,0.0022035555707083806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,3072,0.037979556454552546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,1024,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,256,0.015134221977657743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,3584,0.010509333676762052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,5120,0.016833777228991192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,3584,0.008776889079146916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,7168,0.019928889142142404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,6144,0.13986489507887098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,5120,0.019872888922691345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,256,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,768,0.01717866626050737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,1024,0.007684444387753804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,8192,0.1723173326916165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,512,0.002519999941190084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,16384,0.02815999918513828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,1536,0.03266755408710904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,10240,0.011730666789743634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,3584,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,7168,0.024364444944593642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,10240,0.051367110676235624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,2048,0.008972444468074376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,2048,0.009117333425415887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,768,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,8192,0.02497422198454539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,5120,0.007351111206743453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,8192,0.02639377779430813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,2048,0.04134755664401584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,2560,0.02091911103990343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,3072,0.04103911254141066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,3072,0.009279111193286048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,16384,0.02407022151682112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,3072,0.0073315559162033936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,32,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,2560,0.030681778987248737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,16384,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,32,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,16384,0.016152888536453247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,3584,0.03422844409942627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,12288,0.01424088908566369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,2560,0.017108443710539076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,12288,0.014808000789748298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,2560,0.006953777538405524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,10240,0.027471999327341717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,768,0.007285333342022366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,64,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,5120,0.007446222007274628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,10240,0.01309333327743742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,7168,0.11999911732143825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,1536,0.013410666750537025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,256,0.01705333259370592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,65536,1.5983929104275172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,4096,0.00831200016869439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,128,0.019241778386963736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,3072,0.03993955585691664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,3072,0.00904888908068339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,4096,0.05057511064741346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,7168,0.013591999808947245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,4096,0.009301332963837517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,6144,0.012813333008024426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,512,0.003903999924659729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,1024,0.02604355580276913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,1024,0.0032586666444937387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,65536,0.1480284399456448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,4096,0.02482755482196808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,1024,0.008657777474986183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,16384,0.34946489334106445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,7168,0.009923555784755284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,256,0.0028880000528362063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,1536,0.008291555775536431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,256,0.0021804444905784395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,256,0.0052471111218134565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,4096,0.16074932946099177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,7168,0.023030221462249756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,16384,0.012738666600651212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,10240,0.010689777632554373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,7168,0.03328266739845276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,5120,0.04227200150489807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,64,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,64,0.0028408887899584244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,1536,0.01496444476975335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,2560,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,1024,0.003522666792074839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,2560,0.00740533322095871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,5120,0.01478488908873664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,1024,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,768,0.01276266657643848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,4096,0.012636444634861417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,1024,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,2560,0.03346044487423367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,10240,0.01681511104106903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,65536,0.08852355347739325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,4096,0.07995911439259847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,7168,0.016032889485359192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,256,0.006320000108745363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,1536,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,6144,0.6486079957750108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,768,0.017858665850427415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,64,0.002522666628162066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,256,0.022172444396548804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,8192,0.01204622205760744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,1536,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,5120,0.009353777600659264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,12288,0.03126400046878391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,2560,0.08748710817760891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,1024,0.01817599932352702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,16384,0.016846223009957206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,6144,0.01235466698805491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,5120,0.024960888756646052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,1024,0.29658399687872994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,16384,0.42922221289740664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,16384,0.025506666964954797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,6144,0.9956240124172635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,768,0.005964444329341252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,8192,0.01478666729397244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,16384,0.043903999858432345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,7168,0.0362755556901296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,65536,0.14877155092027453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,1536,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,10240,0.02769600020514594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,64,0.002199999988079071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,128,0.002838222309947014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,64,0.001826666709449556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,2048,0.01752444439464145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,16384,0.02035999960369534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,256,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,768,0.002536000063021978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,768,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,16384,0.014398222168286642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,2560,0.004868444469239977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,2560,0.005370666583379109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,5120,0.024862221545643274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,32,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,768,0.0041457778877682155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,5120,0.02697155541843838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,2048,0.0047395556337303585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,10240,0.05522133244408501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,12288,0.038408888710869685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,3584,0.023041778140597876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,65536,0.12346666389041477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,16384,0.38637423515319824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,768,0.006224888894293044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,768,0.0058400001790788435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,1536,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,512,0.002385777731736501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,2560,0.00723555518521203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,4096,0.025019556283950806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,3584,0.06780266761779785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,8192,0.010703999963071613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,7168,0.010034666293197209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,5120,0.17581510543823242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,8192,0.011003555523024665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,12288,0.03418933351834615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,12288,0.46093246671888566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,1024,0.01310222182008955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,1024,0.018732445107565988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,768,0.023744889431529578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,512,0.002472888885272874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,512,0.01237511138121287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,2560,0.06406044297748141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,8192,0.1531511147816976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,7168,0.029306666718588933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,2048,0.006681777950790193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,128,0.044461333089404635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,256,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,16384,0.044455111026763916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,16384,0.03715377714898851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,5120,0.059282667107052274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,8192,0.21954666243659127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,6144,0.011023110813564725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,8192,0.11314933829837376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,128,0.002833777831660377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,2560,0.009186666872766282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,2048,0.004861333303981357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,10240,0.10603911346859402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,8192,0.01408088869518704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,10240,0.01203555530971951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,128,0.01036266651418474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,6144,0.03522933191723294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,12288,0.07088444630304973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,2560,0.004924444274769889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,6144,0.08655733532375759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,256,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,16384,0.01240355521440506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,128,0.0073626670572492815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,2048,0.038336889611350164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,10240,0.015692444311247934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,4096,0.010425777898894416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,1536,0.005576888720194499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,2048,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,1536,0.04412800073623657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,8192,0.03651555710368686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,65536,0.009463110731707679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,3584,0.009526222116417354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,256,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,768,0.002623999905255106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,64,0.0041457778877682155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,1536,0.14848089218139648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,7168,0.010032888915803697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,10240,0.028039111031426325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,65536,0.2397520012325711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,12288,0.013413333230548434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,128,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,768,0.00619911112719112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,8192,0.016229333149062265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,4096,0.007708444363541073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,64,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,2048,0.01757066614098019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,65536,8.42537350124783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,512,0.027664000789324444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,64,0.009307555854320526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,256,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,7168,0.010439111126793755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,5120,0.012072888513406118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,10240,0.025383111503389146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,64,0.007640889121426477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,2048,0.006693333387374878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,3584,0.008294222255547842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,32,0.006952889263629913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,256,0.0069360000391801195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,768,0.0031644445326593188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,512,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,128,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,8192,0.013388444152143268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,128,0.010025777750545079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,32,0.001818666648533609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,10240,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,2048,0.01924000018172794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,512,0.00793688909875022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,4096,0.0143857780430052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,2560,0.015493333339691162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,2048,0.061195552349090576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,3584,0.007333333293596904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,16384,0.02540622154871623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,4096,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,1024,0.0063182223174307085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,2048,0.008983111215962304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,6144,0.009034666750166152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,16384,0.03095910946528117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,2048,0.02240266733699375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,1536,0.06581422355439928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,256,0.014800000521871777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,1536,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,10240,0.012065778175989786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,768,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,2560,0.020319999919997323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,768,0.005693333430422678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,768,0.013773333695199756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,10240,0.009045333498054082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,16384,0.03649955656793382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,3584,0.09018133083979289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,4096,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,512,0.0028320000403457214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,64,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,5120,0.008971555365456475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,12288,0.00998488896422916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,65536,0.3241537676917182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,3584,0.008933333059151968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,2048,0.01106844428512785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,12288,0.04277866747644213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,2048,0.004521777646409141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,5120,0.017522666189405654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,5120,0.026657778355810378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,5120,0.00999111102686988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,256,0.005251555393139522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,256,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,1536,0.016460445192125108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,12288,0.02330400049686432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,6144,0.2916799916161431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,768,0.10531110896004571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,1024,0.0032151111712058387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,64,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,12288,0.03522933191723294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,4096,0.014775110615624322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,65536,0.11400977770487468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,3072,0.007995555798212687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,256,0.040398223532570735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,3072,0.06302311023076375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,768,0.005527110977305307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,65536,0.09043200148476495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,1024,0.013769778112570444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,16384,0.019900444481107924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,1536,0.010566222170988718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,6144,0.025815111067559984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,64,0.0021564445147911706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,256,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,1024,0.020240889655219186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,4096,0.01732711162832048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,512,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,2048,0.028693334923850164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,10240,0.08999733130137126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,16384,0.17761600017547607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,256,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,7168,0.04787910977999369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,12288,0.02813777658674452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,3584,0.009160888691743216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,512,0.004037333445416557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,3584,0.008201777935028076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,65536,0.39455289310879177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,2560,0.009848888549539778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,1536,0.02163644466135237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,65536,0.13032266828748915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,768,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,10240,0.02083200050724877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,256,0.011113777756690979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,768,0.007385777930418651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,512,0.0058764442801475525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,2048,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,6144,0.018339556124475267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,512,0.00924355536699295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,7168,0.010328888893127441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,6144,0.011346666349305047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,3584,0.04250133368704054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,3072,0.020279111133681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,1536,0.023932443724738225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,2560,0.007248889240953658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,65536,0.01013777818944719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,12288,0.10284355613920425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,256,0.015455999308162265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,8192,0.026366222235891554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,6144,0.019984000258975558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,5120,0.02769244379467434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,5120,0.24425954288906523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,64,0.020577778418858845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,768,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,65536,0.35500889354281956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,2048,0.013286222186353473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,5120,0.013380444712109037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,4096,0.0053191110491752625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,2048,0.007243555453088548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,256,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,768,0.024354666471481323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,16384,0.07206577724880643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,16384,0.025794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,10240,0.03052799900372823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,768,0.006952889263629913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,512,0.012806221842765808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,768,0.01757955551147461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,16384,0.04121155540148417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,512,0.013957333233621387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,1024,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,65536,0.05275288886494107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,256,0.0069413334131240845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,4096,0.018181333939234417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,10240,0.027330666780471802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,3072,0.10406044456693862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,65536,0.08158222171995375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,3072,0.2641048961215549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,8192,0.022335110439194575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,2560,0.006940444310506185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,1024,0.00850755555762185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,6144,0.0199297782447603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,12288,0.01940266622437371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,10240,0.03200533323817783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,768,0.012144000165992312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,64,0.00490044429898262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,65536,0.057793776194254555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,2560,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,10240,0.02188355558448368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,1536,0.0033031110134389666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,10240,0.01575911045074463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,7168,0.013796444568369122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,5120,0.008443555898136562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,12288,0.016935111747847665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,65536,0.07489777935875787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,5120,0.00870400004916721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,7168,0.0462026662296719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,1536,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,4096,0.015581332974963717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,12288,0.026224889689021643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,768,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,1536,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,512,0.023665777511066858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,65536,0.15210044384002686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,65536,0.12868444124857584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,2560,0.031361778577168785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,2560,0.017248888810475666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,768,0.012343111137549082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,64,0.0021439999755885866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,10240,0.1779031091266208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,512,0.0052826665341854095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,3584,0.00980977796845966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,65536,0.15073245101504856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,256,0.01484266585773892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,6144,0.01629333363638984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,3584,0.01722400055991279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,4096,0.007608888877762689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,2560,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,2560,0.005203555441564984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,512,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,256,0.006646222124497096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,65536,0.0864337748951382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,256,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,1536,0.012198222180207571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,1024,0.02777422302299076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,16384,0.025472889343897503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,4096,0.008611555728647444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,5120,0.04689688815010918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,65536,0.09286044703589545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,7168,0.030572444200515747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,4096,0.02182666626241472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,5120,0.01019200020366245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,64,0.0021466666625605687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,65536,0.04686399963166979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,6144,0.005277333160241445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,6144,0.022989332675933838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,128,0.0052666668262746595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,7168,0.266908433702257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,3584,0.007654222349325816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,3072,0.018959111637539335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,512,0.014449778530332776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,7168,0.014110222458839417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,3584,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,12288,0.00489688871635331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,16384,0.03902577691608005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,3584,0.013609777722093793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,4096,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,6144,0.009644444617960189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,768,0.01742844449149238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,512,0.0025351111673646504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,8192,0.04890577660666572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,7168,0.031015998787350122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,10240,0.012437333663304647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,7168,0.010431999961535135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,7168,0.01037333326207267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,8192,0.013282666603724161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,8192,0.02971200148264567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,3584,0.0075911109646161394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,4096,0.011356444822417365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,512,0.013411555853154926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,2048,0.04518489042917887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,65536,0.32223378287421334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,10240,0.017498667041460674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,6144,0.013288000391589271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,4096,0.10691110955344306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,16384,0.1535795529683431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,5120,0.04663999875386556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,64,0.006633777585294511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,256,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,128,0.0018977777411540349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,1536,0.019136000010702345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,64,0.025123554799291823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,768,0.004874666531880696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,768,0.0025253333151340485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,10240,0.03832177652253045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,65536,0.0628968874613444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,256,0.015470221638679504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,12288,0.04228355487187704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,2048,0.014351111319330005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,8192,0.015793777174419828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,6144,0.05494311120775011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,6144,0.019504888190163504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,256,0.004325333154863781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,128,0.006833777659469181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,256,0.009497777455382878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,128,0.002128888956374592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,3072,0.007877333296669854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,64,0.0052471111218134565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,7168,0.012307555311255984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,8192,0.015079110860824585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,10240,0.141576886177063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,7168,0.010023111270533668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,12288,0.017857778403494094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,768,0.015141333142916361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,6144,0.018535999788178336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,6144,0.004583111239804162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,10240,0.051207111941443555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,65536,0.06691110796398587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,128,0.0236826671494378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,768,0.015126221709781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,3584,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,7168,0.062000003125932485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,4096,0.009874666730562845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,768,0.0031795555518733132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,3584,0.04565599891874525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,3584,0.02126577827665541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,64,0.012684444586435953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,2048,0.13618132803175184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,768,0.015803555647532146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,3072,0.01683822274208069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,1536,0.020982222424613103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,4096,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,2048,0.004247111164861255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,1536,0.015470221638679504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,16384,0.04521422253714668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,8192,0.01945600079165565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,2560,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,128,0.016169778174824186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,32,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,2560,0.0042204442951414315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,2048,0.025656888882319134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,8192,0.01016711112525728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,8192,0.022816000713242426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,16384,0.23428267902798125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,12288,0.07727822330262926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,1024,0.020605333977275424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,16384,0.016151999433835346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,6144,0.01814933286772834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,65536,4.73344251844618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,128,0.014787554740905762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,768,0.018558222386572096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,1024,0.02274399995803833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,768,0.007975111405054728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,64,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,10240,0.014820444915029736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,2048,0.06904888815350003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,512,0.014110222458839417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,65536,0.18715200159284803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,16384,0.016511110795868766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,65536,0.06262133518854777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,3072,0.09020444419648911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,12288,0.03668533431159125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,2048,0.009022222624884712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,1536,0.004544888933499654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,2048,0.006507555643717448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,512,0.002463999928699599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,8192,0.03940800163480971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,5120,0.014390221900410123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,768,0.003301333429084884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,4096,0.08229066928227742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,128,0.015488889482286243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,2560,0.00999022192425198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,3584,0.012714666624863943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,1536,0.007322666545708974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,8192,0.011716444459226398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,7168,0.010355555348926121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,128,0.00217333332531982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,65536,0.10266577535205418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,12288,0.043623109658559166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,65536,0.9638524585300021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,2048,0.015374221735530429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,4096,0.014150222142537435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,3072,0.007576000359323289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,3072,0.01592088904645708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,4096,0.04714666803677877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,2048,0.011961778004964193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,512,0.06083999739752876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,128,0.005072888814740711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,4096,0.12251110871632893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,1024,0.009664889011118147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,8192,0.014306666122542487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,32,0.021563554803530376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,5120,0.024175999893082514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,4096,0.004554666578769684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,6144,0.014330666926172046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,32,0.008989333278603023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,2560,0.019662222928471036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,2560,0.019786667492654588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,768,0.01792622274822659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,3072,0.018253333038753934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,2560,0.009775111244784461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,65536,0.08208622535069783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,2048,0.013413333230548434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,8192,0.017484444710943434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,65536,0.841203530629476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,7168,0.010324444207880232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,3584,0.011005333728260465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,768,0.0038337777886125776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,2560,0.008330666356616551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,10240,0.03694577680693732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,5120,0.0535733335547977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,10240,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,6144,0.12102222442626953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,6144,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,3584,0.09638666444354588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,1536,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,8192,0.020451555649439495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,128,0.008319999608728621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,12288,0.10778933101230198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,5120,0.01036711119943195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,64,0.002152888932161861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,10240,0.020249777370029025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,16384,0.04260977771547106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,256,0.015139554937680563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,65536,0.06323644187715319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,12288,0.018364444375038147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,128,0.005620444400442972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,2560,0.03050311075316535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,16384,0.043707556194729276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,3072,0.03149866726663377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,3584,0.018177777528762817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,10240,0.016787555482652452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,64,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,3584,0.020084444019529555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,64,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,768,0.004081777814361784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,32,0.0021137777302000257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,5120,0.0136808885468377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,1536,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,256,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,12288,0.017116443978415597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,3072,0.008115555677149031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,128,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,8192,0.021874666213989258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,768,0.0031822222388452957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,4096,0.008057778080304464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,10240,0.08097066481908162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,16384,0.0551093320051829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,2560,0.014487110906177096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,8192,0.024646222591400146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,2048,0.011896889242861005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,7168,0.1616764465967814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,5120,0.004540444248252445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,6144,0.006705777926577463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,10240,0.2026826673083835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,256,0.003170666595300039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,512,0.006626666833957036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,12288,0.10353066523869832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,16384,0.01791288952032725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,2560,0.019883554842736986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,1536,0.024596444434589807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,2048,0.01442399952146742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,768,0.005950222412745158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,6144,0.01422933406300015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,3584,0.012383999923865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,16384,0.10770844750934178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,5120,0.009361777868535783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,128,0.007600888609886169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,7168,0.03377866744995117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,65536,0.07102400064468384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,2048,0.020064888728989493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,768,0.002855111120475663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,8192,0.19420711199442545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,4096,0.010281778044170803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,3072,0.004879999905824661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,4096,0.07154577970504761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,16384,0.02019377715057797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,65536,0.16545244057973227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,10240,0.030982222821977403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,32,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,4096,0.021124444074100916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,1536,0.017216889394654166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,128,0.008645333349704742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,2560,0.22256800863477919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,4096,0.020752888586786058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,8192,0.010666666759385003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,1024,0.025036444266637165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,12288,0.013443555268976422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,2560,0.005097777893145879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,3072,0.019816888703240287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,4096,0.008307555483447181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,5120,0.03257688879966736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,12288,0.04255022274123298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,3072,0.07031110922495525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,65536,0.08900355630450779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,3072,0.006496888895829518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,2560,0.046262221203909985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,12288,0.01883288886811998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,65536,0.06886933247248332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,6144,0.037639998727374606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,10240,0.15576355987124973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,3072,0.01036977767944336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,768,0.03161333335770501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,7168,0.03379822108480666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,256,0.004941333499219683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,512,0.008356444537639618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,10240,0.1139404508802626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,8192,0.028347555134031508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,32,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,6144,0.009348444640636444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,7168,0.022075555390781824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,3072,0.0094355551732911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,6144,0.041025777657826744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,1024,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,6144,0.010247111320495605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,32,0.001790222194459703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,8192,0.012068444656001197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,16384,0.2926959991455078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,8192,0.011400000088744693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,10240,0.20072000556521943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,12288,0.02576444380813175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,3072,0.02979466650221083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,65536,0.07657155725691053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,65536,0.058125330342186816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,12288,0.13533688916100398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,1536,0.01367377738157908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,128,0.0021679999513758556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,2048,0.00658044425977601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,65536,0.12647377120123968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,2560,0.0343777769141727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,2560,0.10335555341508652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,3584,0.07803999715381198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,8192,0.029158221350775823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,16384,0.013519110778967539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,16384,0.018221333622932434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,7168,0.02331377731429206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,1024,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,128,0.0021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,16384,0.07805777920616998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,6144,0.021378666162490845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,128,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,768,0.004173333446184794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,65536,0.07142933209737141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,3584,0.01815733313560486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,8192,0.013749333719412485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,2048,0.01581333412064446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,1536,0.012063110868136087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,1536,0.0038453332251972626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,2560,0.00699733363257514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,5120,0.09491999944051106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,7168,0.010031999813185798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,7168,0.015841777125994366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,7168,0.017048888736300997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,768,0.005752000129885144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,3584,0.07193955447938707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,1536,0.006877333339717653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,32,0.0069146665434042616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,2560,0.08553600311279297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,6144,0.007334222396214803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,12288,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,65536,0.06956266694598727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,512,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,12288,0.011747555600272285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,4096,0.023980443676312763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,768,0.007658667034573025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,5120,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,4096,0.008656000097592672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,10240,0.011748444702890186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,10240,0.04203733470704821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,7168,0.014344000154071383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,256,0.0038417776425679526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,256,0.002557333351837264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,64,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,5120,0.010423999693658618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,128,0.03025777803526984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,3584,0.007992000215583378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,65536,0.06606488757663302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,2560,0.0069679998689227635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,3584,0.0262773334980011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,65536,0.22320887777540419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,128,0.0021688888470331826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,1024,0.010710222025712332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,65536,6.165690528021918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,3584,0.011352889239788055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,3072,0.019716444942686293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,1024,0.00868711123863856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,3584,0.03056710958480835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,12288,0.04307022359636095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,8192,0.01551466683546702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,768,0.021938666701316833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,4096,0.009654222263230218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,256,0.002528000002106031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,3072,0.11852978335486518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,65536,0.13056888845231798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,768,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,10240,0.07530311081144545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,128,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,768,0.012812444733248817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,1024,0.016792888442675274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,4096,0.00795733349190818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,2048,0.02467022173934513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,6144,0.018225777480337355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,16384,0.06126133600870768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,3584,0.01240711079703437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,16384,0.17912622292836508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,64,0.002509333400262727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,7168,0.01000355515215132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,2560,0.008979555633332994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,32,0.004863111095296012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,6144,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,12288,0.02980355421702067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,5120,0.46402756373087567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,5120,0.014352000421947904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,12288,0.1312675608528985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,512,0.013925333817799887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,64,0.016840888394249808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,512,0.006322666588756774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,16384,0.052858667241202466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,1536,0.0038684445122877755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,64,0.0017964444640609955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,8192,0.00904622260067198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,3072,0.025786666406525507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,2560,0.008071111308203803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,8192,0.07153511047363281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,2560,0.020058666666348774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,8192,0.020624000165197585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,2560,0.016557332542207506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,1536,0.008018666671382057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,7168,0.009656888743241629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,8192,0.030454221698972914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,256,0.0028106667515304354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,16384,0.026308443811204698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,2560,0.006635555376609166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,3584,0.007285333342022366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,6144,0.017902221944597032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,16384,0.06170755624771118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,2560,0.015522667103343539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,65536,0.06725333134333293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,12288,0.0573635564910041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,128,0.02731288969516754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,10240,0.04339822133382162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,2560,0.02197333342499203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,1024,0.005647111270162795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,7168,0.013497778111033969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,512,0.006007110906971826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,128,0.005031110925806893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,5120,0.009024000002278222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,12288,0.015042665931913586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,128,0.014471999473041959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,2560,0.015426667200194465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,16384,0.04197955462667677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,16384,0.04910666743914286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,6144,0.022454222043355305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,8192,0.02495644489924113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,4096,0.024266666836208765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,64,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,8192,0.05093600021468269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,32,0.0018444444156355327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,65536,0.2169733312394884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,5120,0.008346666892369589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,4096,0.02509333358870612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,7168,0.0214684440029992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,12288,0.023725334140989516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,64,0.002895999906791581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,12288,0.029425776667065088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,65536,0.0897173351711697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,10240,0.02628266645802392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,128,0.014456889695591398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,256,0.002842666581273079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,5120,0.014398222168286642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,2560,0.008014221986134848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,32,0.005240889059172736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,7168,0.02294755478700002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,65536,0.11400443977779812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,16384,0.03218844532966614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,2048,0.016174222032229107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,7168,0.02020355562369029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,6144,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,2048,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,2048,0.014954666296641031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,128,0.01447111037042406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,2048,0.07093689176771376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,128,0.014858666393491956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,12288,0.1032702194319831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,5120,0.02756799923049079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,2560,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,2560,0.007244444555706448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,8192,0.015777778294351365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,768,0.003141333245568805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,10240,0.01921688848071628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,10240,0.04297511114014519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,32,0.0052639999323421055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,7168,0.012031111452314587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,5120,0.031155556440353394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,256,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,512,0.005330666485759947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,8192,0.010674667027261523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,128,0.04161422120200263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,7168,0.014935111006100973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,512,0.008954666554927826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,64,0.008283555507659912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,64,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,2048,0.05023733443684048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,2048,0.019596444235907663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,3072,0.008919110728634728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,7168,0.018830221560266282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,64,0.002069333361254798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,10240,0.012427555190192329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,2560,0.004954666727119022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,1024,0.007652444144090016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,2048,0.012011555333932241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,4096,0.007898666792445714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,2560,0.02737422287464142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,12288,0.013845333622561561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,512,0.026964444253179762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,6144,0.015036443869272867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,6144,0.010058666268984476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,768,0.01682488951418135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,768,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,6144,0.013117333253224691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,4096,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,768,0.0064106666379504735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,8192,0.016794666647911072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,10240,0.06656889120737712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,1536,0.012741333080662621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,32,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,7168,0.01402844488620758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,4096,0.016488000750541687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,8192,0.04766844378577339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,3584,0.009160888691743216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,3584,0.00775111135509279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,2560,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,10240,0.03237244486808777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,2048,0.031166222360399034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,6144,0.24820621808369955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,10240,0.028528001573350694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,2048,0.04119022356139289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,128,0.0021768889079491296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,128,0.006265777680608962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,4096,0.02064533366097344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,5120,0.02104444470670488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,5120,0.008981333010726506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,1024,0.009709333380063375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,5120,0.024079999989933435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,5120,0.01200711147652732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,16384,0.17001867294311523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,4096,0.014497778481907315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,64,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,2560,0.06565244330300225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,2048,0.006528888725572162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,12288,0.04294311006863912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,6144,0.006729777902364731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,256,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,5120,0.02068444424205356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,6144,0.009555555880069733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,6144,0.06307022439108954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,2048,0.007409777906205919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,5120,0.026107556290096704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,12288,0.6452337900797526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,1536,0.018064889642927382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,512,0.08803555700514053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,65536,0.10003644227981567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,65536,0.20984444353315565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,2048,0.01925155520439148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,65536,0.03667733404371474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,7168,0.011477332976129321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,64,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,64,0.00216977774269051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,5120,0.03611466619703505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,10240,0.014281777871979607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,12288,0.01756088932355245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,5120,0.01016622202263938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,12288,0.04415377643373278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,128,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,3072,0.01035644445154402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,5120,0.021700445148679946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,10240,0.030768000417285498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,12288,0.019409777389632333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,10240,0.10728622145122951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,32,0.02206755512290531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,3072,0.03048622277047899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,768,0.024343111448817786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,3072,0.01942933268017239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,768,0.011420444481902652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,256,0.0021822222818930945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,1024,0.006222222414281633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,128,0.0021119999388853707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,64,0.005601777798599667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,1536,0.00793511089351442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,1024,0.004872888740566042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,3072,0.007979555262459649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,2048,0.01718933383623759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,512,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,2560,0.006976000136799282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,3584,0.0347671111424764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,7168,0.0052595556610160405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,12288,0.14891733063591853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,10240,0.05334577626652188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,256,0.020963556236690942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,128,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,65536,0.20541155338287354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,4096,0.26142933633592397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,768,0.016505777835845947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,8192,0.026182222697469924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,7168,0.020834666159417894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,4096,0.03907466597027249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,4096,0.009335999687512716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,3072,0.11152533690134685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,12288,0.02821422285503811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,12288,0.04618488748868307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,12288,0.021615111165576514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,3072,0.009755555954244401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,1024,0.005984888722499211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,768,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,3584,0.01846222248342302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,5120,0.01015022231472863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,3072,0.09883111053042942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,6144,0.022618666291236877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,16384,0.020810667011472914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,1536,0.01015911085738076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,3584,0.016516443755891588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,3072,0.013277333643701343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,4096,0.05054222212897407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,32,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,2560,0.005193777796294955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,8192,0.01538755496342977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,128,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,4096,0.22550400098164877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,512,0.033965332640541926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,6144,0.018537777993414137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,2560,0.02092622220516205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,8192,0.020569778150982328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,64,0.005327111317051782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,10240,0.0473653342988756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,256,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,8192,0.011699555648697747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,2048,0.010828444527255164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,4096,0.017895999881956313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,8192,0.19793244202931723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,2560,0.011977777712874942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,65536,0.3566106690300836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,12288,0.06473866436216566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,12288,0.028103109863069322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,1024,0.0035537779331207275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,512,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,5120,0.004842666702138053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,512,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,64,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,6144,0.023331556055280898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,512,0.0041946665280395085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,3584,0.00813066628244188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,256,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,512,0.013811555173661975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,3584,0.01089866707722346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,2048,0.007847111258241866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,7168,0.05613333649105496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,768,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,1536,0.028971556160185073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,6144,0.005955555372767978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,6144,0.019190222024917603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,1536,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,10240,0.011687999798191918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,3072,0.011755555868148804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,6144,0.01903199983967675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,768,0.004396444393528832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,256,0.006250666661394968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,16384,0.017895999881956313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,7168,0.011896889242861005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,2048,0.008319999608728621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,3072,0.011051555474599203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,2048,0.00656355544924736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,256,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,6144,0.008982222113344405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,2560,0.012056888805495368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,3584,0.004856889032655292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,10240,0.004925333377387789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,4096,0.009193778038024902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,12288,0.02163733376397027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,128,0.03433955709139506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,768,0.041167110204696655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,32,0.020591111646758187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,3584,0.005323555734422471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,512,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,8192,0.025013332565625507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,3072,0.021686222818162706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,16384,0.09313066800435384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,16384,0.03865955604447259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,1024,0.0041306668685542215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,16384,0.00959822204377916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,4096,0.2235813405778673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,128,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,1536,0.04297777679231432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,2048,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,128,0.02885599931081136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,1024,0.03797511259714762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,10240,0.060807108879089355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,16384,0.10014488961961533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,10240,0.01257155587275823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,128,0.003170666595300039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,3584,0.008105778031879002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,7168,0.01073066641887029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,10240,0.012062222593360476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,10240,0.020580444071027968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,4096,0.010077333284748925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,65536,0.2516453266143799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,65536,0.04684355523851183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,1536,0.006561777657932705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,128,0.005407111098368962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,8192,0.016529776983790927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,1536,0.008621333373917473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,65536,0.0647982226477729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,8192,0.1926248868306478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,7168,0.05491822295718723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,2048,0.02186577849917942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,256,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,2560,0.007306666837798224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,4096,0.027299554811583623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,12288,0.013815999858909182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,1536,0.0064026667839950984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,256,0.025027554896142747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,1536,0.010072888599501716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,7168,0.035439110464519925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,10240,0.008652444514963362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,8192,0.016324443949593436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,2560,0.004583999928500918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,6144,0.12560799386766222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,1536,0.014175110393100314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,4096,0.03179288903872172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,65536,0.17626489533318412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,4096,0.020233778489960563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,3072,0.027858667903476294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,1536,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,3584,0.011029333704047732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,4096,0.008766222331258986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,7168,0.010828444527255164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,2560,0.04257333278656006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,12288,0.01570311188697815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,3584,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,2560,0.09014755487442017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,1024,0.0038568890757030914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,2560,0.011015111373530494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,128,0.015471110741297403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,12288,0.013760888742076026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,12288,0.03175466590457492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,768,0.005746666755941179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,2560,0.009307555854320526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,6144,0.009223110973834991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,12288,0.05095377895567152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,12288,0.18631466229756674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,1024,0.016859556237856548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,128,0.002163555473089218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,12288,0.03927200039227804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,10240,0.01623288955953386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,65536,0.15622756216261122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,128,0.0192239996459749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,3072,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,6144,0.009647111097971598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,1024,0.01686311099264357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,8192,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,1536,0.0059164443777667145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,1024,0.0127368892232577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,1536,0.018206222189797294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,256,0.2571049001481798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,32,0.002559111143151919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,3584,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,7168,0.010330666270520952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,1024,0.00654666663871871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,6144,0.009010666774378883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,3072,0.1227217780219184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,2048,0.009351111120647853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,4096,0.010951110886202918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,6144,0.02808622188038296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,3584,0.007345777418878343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,3584,0.00798044436507755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,2048,0.013122667041089801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,5120,0.016858667135238647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,8192,0.02821422285503811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,10240,0.012110222544935016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,64,0.004538666870858935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,7168,0.011032889286677042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,128,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,3072,0.007687110867765214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,1536,0.026424888107511733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,5120,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,768,0.0028693332440323303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,65536,0.07932888799243502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,6144,0.011694221860832639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,256,0.013644444445768992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,512,0.006667555620272954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,256,0.007965332931942409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,32,0.002855111120475663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,10240,0.025133333272404138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,1536,0.016010666886965435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,16384,0.08831466568840875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,7168,0.02796799937884013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,2560,0.0299911101659139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,10240,0.039164443810780845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,2048,0.0064444442590077715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,6144,0.05436444282531738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,7168,0.014188443620999655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,1024,0.006298666612969504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,65536,0.07275378041797213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,12288,0.10674222310384114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,5120,0.011000889043013254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,6144,0.026722666290071275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,3584,0.030990223089853924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,2048,0.013047999805874295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,65536,0.15350843800438776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,6144,0.01071733319097095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,1536,0.01179377817445331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,3584,0.008009778128729926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,2048,0.008440888590282863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,16384,0.2596728801727295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,5120,0.016677333248986136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,3584,0.04381155636575487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,256,0.06495022111468844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,1536,0.04448711209826999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,7168,0.03685599896642897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,768,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,2560,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,10240,0.08008355564541288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,3072,0.20678222179412842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,128,0.014487110906177096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,3584,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,5120,0.013967111706733704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,5120,0.03029866682158576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,8192,0.012668444050682915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,2048,0.011055111057228513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,2048,0.032255109813478254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,512,0.005547555370463266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,8192,0.02676533328162299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,12288,0.013439111411571503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,1024,0.027639999985694885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,1536,0.023434667123688593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,6144,0.012616000241703458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,512,0.024508444799317256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,128,0.0025128889828920364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,512,0.015806222955385845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,3584,0.0789440009329054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,128,0.017104888955752056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,6144,0.07134222322040133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,5120,0.008664000365469191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,7168,0.014753777119848462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,12288,0.04582133226924472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,2560,0.006728000111050076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,12288,0.0310204459561242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,1024,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,64,0.05728799766964383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,8192,0.01203555530971951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,4096,0.020085333122147452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,7168,0.034456888834635414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,512,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,512,0.0041724443435668945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,5120,0.01722044414944119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,2560,0.009638222555319468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,8192,0.007310222420427535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,5120,0.012808889150619507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,6144,0.04008177916208903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,128,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,2048,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,6144,0.011400000088744693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,768,0.03882222374280294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,64,0.005535999933878581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,3584,0.013893332746293811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,12288,0.049477332168155246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,8192,0.18611466884613037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,6144,0.009387555221716562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,10240,0.021687111920780603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,4096,0.00795555528667238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,5120,0.02075466679202186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,32,0.0018728888697094387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,3072,0.006400888992680444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,32,0.002157333410448498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,16384,0.021176000436147053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,5120,0.023716444770495098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,65536,0.14051466517978245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,768,0.009853333234786987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,10240,0.02662311163213518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,10240,0.03345866666899787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,1536,0.02088800072669983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,6144,0.011766222616036734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,10240,0.0989066693517897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,5120,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,7168,0.018022222651375663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,16384,0.023201778531074524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,4096,0.004544888933499654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,1024,0.013786666923099093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,65536,0.008994667066468133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,32,0.007611555357774098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,768,0.015480889214409722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,512,0.008969777988062965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,128,0.005958222266700532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,32,0.0063075555695427795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,4096,0.02180444366402096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,1536,0.031453334622912936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,128,0.014779556128713818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,256,0.0022026666750510535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,6144,0.010671111444632212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,8192,0.027640889088312786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,4096,0.03984444340070089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,6144,0.018574222922325134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,12288,0.022843556271659002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,10240,0.03823644585079617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,16384,0.037866665257347956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,10240,0.3489217758178711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,2048,0.004554666578769684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,2048,0.009827555881606208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,3584,0.1736604505115085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,256,0.2046391169230143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,4096,0.02694844371742672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,512,0.0032133333798911837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,1024,0.006903111106819577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,5120,0.03905866543451945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,4096,0.008904000123341879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,65536,0.09410844246546428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,3584,0.08312355809741549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,4096,0.02522666586769952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,65536,0.009633777870072259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,65536,0.0659360024664137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,1536,0.006215999937719769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,512,0.013376000026861826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,3584,0.027848889430363972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,4096,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,7168,0.028035554620954726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,2048,0.06482844220267402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,256,0.007510222494602203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,3584,1.2901084687974718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,32,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,7168,5.328325483534071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,64,0.00795999997191959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,768,0.010022222167915767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,65536,0.08091200060314603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,64,0.0035128887328836652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,7168,0.014828445182906257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,768,0.017836444907718234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,256,0.02749155627356635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,5120,0.049012445741229586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,2048,0.018536888890796237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,256,0.022500443789694045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,4096,0.008738666772842407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,256,0.029905776182810467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,4096,0.010460444622569613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,10240,0.09394222497940063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,768,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,1536,0.017462222112549674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,5120,0.012631999949614206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,64,0.0069324444565508105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,768,0.006215111249023014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,128,0.0022311111291249595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,128,0.002157333410448498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,7168,0.04244177871280246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,10240,0.01706755492422316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,65536,0.7249253590901693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,2560,0.01883555617597368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,256,0.015127999915017022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,64,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,512,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,12288,0.0184906671444575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,4096,0.028502222564485338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,12288,0.019273777802785236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,8192,0.02010222276051839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,2048,0.0073626670572492815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,7168,0.18613510661655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,12288,0.08797333637873332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,10240,0.21837333838144937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,5120,0.01740444368786282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,1024,0.0028524444335036804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,3584,0.03464533223046197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,10240,0.2858320077260335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,7168,0.01035199976629681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,5120,0.037103109889560275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,6144,0.024711999628278945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,7168,0.005389333185222413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,1024,0.010622222390439777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,256,0.0026808888134029177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,16384,0.046446220742331616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,128,0.006697777658700943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,256,0.004911111046870549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,128,0.0083137775460879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,64,0.006718222051858902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,1024,0.1455671124988132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,64,0.0018399999373488957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,65536,1.9040951199001734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,64,0.0069875551594628235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,65536,0.07548089159859551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,7168,0.0110871113008923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,128,0.028565333949195013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,65536,0.12908621629079184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,32,0.0028942223224374983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,12288,0.005630222045713001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,2560,0.009321778184837764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,2560,0.07605689101748996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,3584,0.007896888587209914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,5120,0.017320000463061862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,3584,0.02984444300333659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,2048,0.011034666664070554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,7168,0.013743110828929476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,2048,0.005633777628342311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,2560,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,1024,0.009305777649084726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,128,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,4096,0.008302222523424361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,3072,0.1475777758492364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,3584,0.007672888537247975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,1536,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,512,0.03371999992264642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,1536,0.006212444355090459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,256,0.005679111099905438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,1536,0.0041582224269707995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,768,0.0026613333159022858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,1536,0.03873599900139703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,128,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,128,0.020241777102152508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,2048,0.009304888546466827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,768,0.015787555111779105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,65536,0.05623110797670153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,1024,0.07755466964509752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,128,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,7168,0.03421155611673991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,3584,0.00664533343580034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,256,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,128,0.008309333688682979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,10240,0.028149333265092637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,12288,0.10005333026250203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,7168,0.010509333676762052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,768,0.005600888695981767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,256,0.01516711049609714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,2560,0.005085333353943295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,12288,0.013569778038395775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,6144,0.009307555854320526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,12288,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,2048,0.006801777829726537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,6144,0.06194666359159681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,7168,0.052654223309622876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,10240,0.022662222385406494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,256,0.016537777251667447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,2048,0.006412444429265128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,512,0.003165333428316646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,4096,0.014166222678290473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,128,0.011382222175598145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,65536,0.2488328880733914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,1536,0.01869244376818339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,256,0.01685599982738495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,1536,0.012705778082211813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,512,0.007003555695215861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,768,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,1536,0.021615111165576514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,64,0.0018586667461527719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,512,0.0031822222388452957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,16384,0.015862221519152325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,2560,0.012345777617560493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,10240,0.012431110772821637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,16384,0.13555288314819336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,512,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,128,0.01443733274936676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,1536,0.006053333481152852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,512,0.015437333120240105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,6144,0.009177777502271865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,768,0.00796977761718962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,768,0.01787911189926995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,5120,0.009367110828558603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,4096,0.014479110638300577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,4096,0.005947555518812603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,3584,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,768,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,768,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,12288,0.014474666780895658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,65536,0.05938666396670871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,128,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,256,0.017844445175594754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,5120,0.021703110800849065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,10240,0.08682755629221599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,256,0.002467555511328909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,128,0.27360889646742076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,1536,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,10240,0.012034667034943899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,768,65536,0.1295200056499905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,12288,0.017864000466134813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,512,0.013756444056828817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,4096,0.007912889122962952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,7168,0.011683555940786997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,5120,0.02659911082850562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,6144,0.01348000019788742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,3072,0.09412355555428399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,10240,0.006250666661394968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,7168,0.03491733471552531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,3584,0.018565333551830716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,3584,0.14356888665093318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,16384,0.02589955594804552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,16384,0.019488000207477145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,6144,0.009008000294367472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,10240,0.015156444576051501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,5120,0.011047110789351992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,768,0.018000889155599806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,64,0.005746666755941179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,10240,0.016679111454221938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,4096,0.009655111365848118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,2560,0.03705866800414191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,4096,0.013630222115251752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,2048,0.011976000335481433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,16384,0.018401778406567044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,6144,0.013398222625255585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,12288,0.03461777832772996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,65536,0.12674310472276476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,4096,0.09852800104353164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,1536,0.01460088955031501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,8192,0.05766133467356364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,7168,0.03581600056754218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,8192,0.03654488921165466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,768,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,512,0.012115555504957834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,7168,0.02436888880199856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,3072,0.012420444852775998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,128,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,2560,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,8192,0.01813333398765988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,12288,0.07153333557976617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,6144,0.02256800068749322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,512,0.0028115556471877625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,2560,0.06456177764468722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,12288,0.03176444437768724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,4096,0.014129777749379476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,2048,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,65536,0.17581421799129912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,768,0.002505777817633417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,65536,0.06086933612823486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,6144,0.14103378189934626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,128,0.001850666685236825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,16384,0.056759999858008496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,2048,0.014819555812411837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,65536,0.09973866409725612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,2048,0.014479110638300577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,7168,0.010386666489972008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,6144,0.011456888582971362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,3584,0.007965332931942409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,512,0.005954666684071223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,3584,0.0075902218619982404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,32,0.01239733315176434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,6144,0.028959999481836956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,8192,0.024999999337726172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,16384,0.02149155570401086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,512,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,2048,0.01963911122745938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,3072,0.12047289477454291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,5120,0.01017777787314521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,65536,0.1879031128353543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,256,0.01717688971095615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,10240,1.042692396375868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,16384,0.7973706457349988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,12288,0.037263999382654824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,65536,0.08651822143130833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,2560,0.008987555901209513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,3584,0.01035288886891471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,6144,0.02107644412252638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,10240,0.11080888907114665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,3584,0.037853333685133196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,128,0.014172444740931192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,64,0.0018115554832749898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,3584,0.007987555530336168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,10240,0.16696266333262125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,768,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,8192,0.009653333160612319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,1024,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,2560,0.00795377790927887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,512,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,3072,0.021753778060277302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,65536,0.6353848775227865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,3584,0.007643555601437886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,2560,0.007309333317809635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,10240,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,1024,0.0032106666929192017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,128,0.01614488826857673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,7168,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,8192,0.011034666664070554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,8192,0.02313066687848833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,12288,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,3072,0.005711110929648082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,256,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,256,0.017845332622528076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,64,0.0018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,6144,0.009703999592198266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,768,0.00889155517021815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,1536,0.013458666702111563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,128,0.0026853332916895547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,256,0.0022115556316243275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,12288,0.046577778127458364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,512,0.0028613333900769553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,2048,0.07027822070651583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,512,0.0147733340660731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,512,0.002467555511328909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,8192,0.025204444924990337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,65536,0.1293813387552897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,128,0.011783999701340994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,65536,0.20991110801696777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,1536,0.0328746669822269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,8192,0.01608088943693373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,1536,0.006259555617968242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,3584,0.10344355636172825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,4096,0.015431111057599386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,4096,0.047157333956824414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,2048,0.037725332710478045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,16384,0.053120000494851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,512,0.0052915554907586836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,3072,0.013067555924256643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,256,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,16384,0.02843466732237074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,6144,0.011196444431940714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,2560,0.030495110485288832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,8192,0.013081777426931592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,1536,0.02676711148685879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,3072,0.007632888853549957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,2560,0.016538666354285348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,16384,0.03789066606097751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,10240,0.011754666765530905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,2560,0.017113778326246474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,10240,0.1118497716055976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,128,0.006647111227114995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,16384,0.04480888777308994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,1024,0.007968000239796108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,768,0.005984888722499211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,3584,0.027375110321574744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,2560,0.008927110996511247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,2560,0.013162666724787818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,7168,0.0199608885579639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,10240,0.063573333952162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,16384,0.14459910657670763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,768,0.0069671107663048645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,512,0.020256888535287645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,3584,0.007648888561460707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,16384,0.02070844504568312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,4096,0.00793333351612091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,3072,0.017280000779363845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,1536,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,768,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,6144,0.01464000013139513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,65536,0.4369662337832981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,12288,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,10240,0.053416000472174756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,1024,0.023652444283167522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,768,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,65536,0.13227910465664333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,64,0.0018595556418100994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,10240,0.07218222485648261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,7168,0.0124177775449223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,3072,0.007592889169851939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,768,0.012755555411179861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,1024,0.04933866527345446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,768,0.002464888824356927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,5120,0.009315555294354757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,5120,0.008656889200210571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,8192,0.05768089161978828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,12288,0.05605777766969469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,2048,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,5120,0.11479643980662028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,512,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,2560,0.010597333312034607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,128,0.015096889601813423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,32,0.006523555351628198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,5120,0.04295288854175144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,5120,0.014527999692493014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,768,0.011406222151385413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,7168,0.03887644410133362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,3584,0.0277839998404185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,6144,0.030140442980660334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,8192,0.010696888797812991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,10240,0.012068444656001197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,3584,0.010693333215183683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,64,0.0024906665914588505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,8192,0.08100355333752103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,12288,1.0519733428955078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,1024,0.016320889194806416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,5120,0.012816000315878125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,2560,0.007674666742483775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,3072,0.007703111403518253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,768,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,32,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,8192,0.018969777557584975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,4096,0.008097777764002482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,256,0.0021475555582178966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,768,0.004084444294373195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,2560,0.08093688885370891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,1536,0.009294222626421187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,256,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,64,0.0035146665241983202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,2560,0.007246221933099959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,128,0.0148062225845125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,256,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,65536,0.06904355684916179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,16384,0.012730666332774691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,3072,0.09362755881415473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,10240,0.030204445123672485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,512,0.01204533295498954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,768,0.0026728889594475427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,1024,0.015781333049138386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,4096,0.03863822089301215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,3584,0.01683111157682207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,12288,0.031204445494545832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,128,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,6144,0.010985777609878115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,512,0.002171555534005165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,12288,0.023569777607917786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,1024,0.18959289126925996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,7168,0.015801777442296345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,10240,0.1823777755101522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,768,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,5120,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,5120,0.06014044417275322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,10240,0.018772444791264005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,3584,0.028978665669759113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,7168,0.018575110369258456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,16384,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,3584,0.022421333524915908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,8192,0.14227022065056696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,8192,0.04612000121010674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,256,0.004908444566859139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,768,0.03947910997602675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,768,0.003148444410827425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,128,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,4096,0.11593243810865615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,64,0.005201777650250329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,10240,0.16232444180382624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,768,0.0026577777332729767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,1024,0.24993599785698783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,1024,0.042338665988710195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,7168,0.007500444849332173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,1024,0.012461333639091916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,2560,0.00868444475862715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,64,0.015182221929232279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,2048,0.008441777692900764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,8192,0.026707554856936138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,6144,0.051856001218159996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,2560,0.02437866727511088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,10240,0.012039999994966718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,5120,0.02163377735349867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,3072,0.0075911109646161394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,5120,0.008640888664457533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,7168,0.013417777915795645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,3584,0.03260711166593764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,12288,0.030797332525253296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,512,0.00655644428398874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,1024,0.0031893334041039148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,16384,0.013399110900031196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,65536,0.0850293305185106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,8192,0.016900445024172466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,8192,0.013233777549531726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,12288,0.27018843756781685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,7168,0.018262222409248352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,2048,0.011494222614500256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,10240,0.03044711218939887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,12288,0.013089777694808112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,12288,0.09840799702538384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,256,0.007963555554548899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,2048,0.07682844665315416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,3584,0.13653955194685194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,2048,0.019478221734364826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,12288,0.019689778486887615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,512,0.010112000008424124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,10240,0.017476444443066914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,4096,0.019959999455346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,7168,0.02462844385041131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,6144,0.009668444593747457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,7168,0.014387556248241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,10240,0.02014400064945221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,2048,0.0314977765083313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,12288,0.01443733274936676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,65536,0.22224177254570854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,128,0.009357333183288574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,512,0.005614222337802251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,7168,0.011387555963463254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,1536,0.0107022225856781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,7168,0.023335999912685815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,1024,0.006605333338181178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,128,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,768,0.015469332536061605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,4096,0.039000888665517174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,3072,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,2048,0.01958400011062622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,6144,0.015496888094478183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,768,0.015825778245925903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,2048,0.0114702218108707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,512,0.005537777725193236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,3072,0.024043555061022442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,3072,0.020928000410397846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,1024,0.00830844458606508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,128,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,1536,0.00674311113026407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,3072,0.007250666618347168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,4096,0.007636444436179266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,768,0.029994666576385498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,768,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,32,0.002491555487116178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,12288,0.016935111747847665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,32,0.0018053334206342697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,12288,0.017473777135213215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,7168,0.023687111006842718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,65536,0.10383466879526775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,768,0.0073315559162033936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,2560,0.030710223648283217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,3584,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,65536,0.019374221563339233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,1536,0.011384888655609555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,2560,0.11203466521369086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,2048,0.027744000156720478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,1024,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,1024,0.007285333342022366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,3072,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,12288,0.02537511123551263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,16384,0.05808977948294746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,1024,0.004628444297446145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,128,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,16384,0.547744009229872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,1536,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,4096,0.012528888881206512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,12288,0.015468445089128284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,128,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,1024,0.02124622298611535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,4096,0.007995555798212687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,512,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,128,0.015099555253982544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,1024,0.0062977779242727495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,65536,0.05532711082034641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,128,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,65536,0.08142577939563327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,12288,0.013088888592190213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,6144,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,768,0.018207111292415194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,6144,0.018264888061417472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,3072,0.007280888656775157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,6144,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,64,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,512,0.025050666597154405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,12288,0.041322665082083807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,16384,0.008251555263996124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,6144,0.011333333121405708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,7168,0.01018666641579734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,7168,0.0052968888647026485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,12288,0.2311404546101888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,512,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,768,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,7168,0.4947226842244466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,128,0.00221422231859631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,128,0.13681688573625353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,7168,0.011049778097205691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,512,0.02474311159716712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,3072,0.0410755541589525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,256,0.0022124445272816554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,4096,0.015138667490747241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,10240,0.032221333848105535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,16384,0.03819466630617777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,65536,0.08409333229064941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,12288,0.040194667047924466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,1536,0.03728177812364366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,65536,0.17231199476453993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,2560,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,16384,0.06407288710276286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,12288,0.04998844530847338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,768,0.019267555740144517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,256,0.01517955627706316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,256,0.002521777732504739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,1024,0.01549600064754486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,8192,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,16384,0.2465724415249295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,8192,0.012352888782819113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,3584,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,4096,0.010699555277824402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,10240,0.007449777589903937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,10240,0.018930666976504855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,3584,0.016710221767425537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,1024,0.006039999839332368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,12288,0.007583111524581909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,16384,0.022290666898091633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,10240,0.04676977793375651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,3072,0.02030311193731096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,128,0.015451555450757345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,128,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,8192,0.0330079992612203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,1024,0.0675377779536777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,6144,0.010300444232092964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,7168,0.04492622282769945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,256,0.01276622215906779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,65536,0.3086408774058024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,256,0.01748533381356133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,2560,0.00519822206762102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,2560,0.021011556188265484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,256,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,5120,0.057851552963256836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,512,0.015456888410780164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,1536,0.003896000070704354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,5120,0.009369778136412302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,1024,0.004613333278232151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,6144,0.020865778128306072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,2560,0.009179555707507664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,3584,0.01808444493346744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,32,0.0028124445428450904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,8192,0.019719110594855416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,3072,0.009875555833180746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,512,0.012624888784355588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,1536,0.004287999951177173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,16384,0.13941333029005262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,128,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,768,0.0058542220956749385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,512,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,3584,0.09636622005038792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,65536,0.0672888888253106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,12288,0.0110871113008923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,1024,0.0064604443808396654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,1024,0.004517333375083076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,256,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,7168,0.03560622202025519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,3584,0.00960799968904919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,5120,0.03730666637420654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,4096,0.17371022701263428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,64,0.0018275556051068837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,1024,0.0034942221310403612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,1536,0.0035093331502543557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,2048,0.008958222137557136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,2560,0.004919111314747068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,1536,0.017624000708262127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,768,0.017904000149832833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,1536,0.009511111511124505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,2560,0.05862666500939263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,2560,0.006684444430801604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,64,0.009687110781669617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,1024,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,10240,0.06012533108393351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,2560,0.020564445190959506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,6144,0.041962666643990405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,1536,0.008959111240175035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,2560,0.00965244405799442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,2048,0.007651555869314406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,7168,0.009687999884287516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,1024,0.019711110326978896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,768,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,7168,0.029206222958034937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,2048,0.006911110960774952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,3584,0.007605333295133378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,3072,0.02035999960369534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,4096,0.015967999895413715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,512,0.009988444546858469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,2048,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,64,0.0018435555199782054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,5120,0.012248888611793518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,256,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,3072,0.014088888963063559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,8192,0.018373333745532565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,16384,0.019926221834288705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,2048,0.005877333382765452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,12288,0.1676888863245646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,6144,0.044287112024095326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,768,0.027427555786238775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,3072,0.010687111152542962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,7168,0.01036711119943195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,1024,0.006675555474228329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,128,0.0021404443929592767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,16384,0.014153778553009033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,6144,0.010346666806273991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,128,0.015088000231319003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,16384,0.11371200614505345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,5120,0.013562666873137156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,3584,0.03796089026663039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,768,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,128,0.03128266665670607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,256,0.023973332511054143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,1024,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,3584,0.038247111770841814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,5120,0.009357333183288574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,512,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,768,0.022298667165968154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,64,0.0018880000958840053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,4096,0.02068444424205356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,128,0.0019191110299693213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,65536,0.01982399986849891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,64,0.0018177777528762817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,2560,0.040919999281565346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,12288,0.03937955697377523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,6144,0.028373334142896865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,8192,0.011408000356621213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,12288,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,768,0.017492444978819955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,32,0.001850666685236825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,128,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,10240,0.012082666986518435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,2048,0.014513777361975776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,10240,0.016473778420024447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,4096,0.008635555704434713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,1536,0.0035386664999855887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,5120,0.016523554921150208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,256,0.013111111190583972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,1536,0.013401777380042605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,12288,0.013912000589900546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,8192,0.017870222528775532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,6144,0.009756444229020013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,2560,0.12290133370293511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,1536,0.006597333484225803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,65536,0.1738711065716214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,256,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,7168,0.012103999654452005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,65536,0.013252444565296173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,3072,0.007244444555706448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,128,0.009558222360081142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,6144,0.010476444330480365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,10240,0.012466666599114736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,8192,0.020241777102152508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,512,0.007616000043021307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,12288,0.03321955601374308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,512,0.009995555712117089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,2048,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,16384,0.05873688724305895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,2048,0.07592888673146565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,12288,0.014133334159851074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,6144,0.03580177823702494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,12288,0.06287555562125312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,64,0.0018880000958840053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,64,0.0021546667234765156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,128,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,12288,0.022966222630606756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,256,0.009457777771684859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,256,0.0031884445084465873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,5120,0.004877333425813251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,512,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,8192,0.014344889256689282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,2560,0.01755555636352963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,4096,0.00793599999613232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,5120,0.01442311041884952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,64,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,2560,0.016878222425778706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,4096,0.02102222210831112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,64,0.0024906665914588505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,7168,0.009129777550697327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,16384,0.021260443660948012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,10240,0.030935999419954088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,768,0.009668444593747457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,2560,0.027685332629415724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,8192,0.010672888822025724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,65536,1.4748986562093098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,1536,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,1024,0.006216889040337668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,128,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,7168,0.009655111365848118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,65536,0.29668090078565806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,5120,0.008978666530715095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,4096,0.01479466590616438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,10240,0.015673778123325773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,64,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,3584,0.005662222289376789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,128,0.005920889063013925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,32,0.0024622221373849446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,4096,0.3412195576561822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,2560,0.00601333338353369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,5120,0.014698666003015308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,4096,0.016075554821226332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,4096,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,2560,0.007309333317809635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,768,0.03428622086842855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,16384,0.025751999682850305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,6144,0.009090666969617208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,65536,0.12650488482581243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,3584,0.014753777119848462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,2560,0.006909333169460297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,2048,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,256,0.01718488832314809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,16384,0.038225776619381376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,3072,0.03594933284653558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,256,0.021288888321982488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,512,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,65536,0.5454959869384766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,8192,0.011020444333553314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,1536,0.04673244555791219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,7168,0.03348533312479655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,1536,0.012845333251688214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,3584,0.01036533299419615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,6144,0.017863111363516915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,2560,0.014774221513006421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,2560,0.005187555733654234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,2048,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,4096,0.00832444429397583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,65536,0.19257866011725533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,3072,0.015856888559129503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,3072,0.004565333326657613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,12288,0.013782222237851886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,512,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,512,0.005613333235184352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,1536,0.0041706665522522396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,768,0.01650399963061015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,8192,0.005601777798599667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,64,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,32,0.0041617775956789655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,3584,0.0389342208703359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,4096,0.007245333658324347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,3072,0.0900791088740031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,256,0.015479111009173922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,65536,0.13016443782382542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,65536,0.058005332946777344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,65536,0.13839288552602133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,1536,0.013125333521101209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,12288,0.1295822196536594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,32,0.003170666595300039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,768,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,65536,0.05883644686804878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,768,0.005975111077229182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,5120,0.027219555444187585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,16384,0.037608888414171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,768,0.006071110980378256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,2560,0.036652445793151855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,512,0.011376000112957425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,5120,0.009132444030708736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,3584,0.02125066684352027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,65536,0.19498044914669463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,1536,0.0041751112374994485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,1024,0.015488889482286243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,16384,0.026525333523750305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,2048,0.010838222172525195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,8192,0.033719112475713096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,2560,0.004683555414279302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,4096,0.013386666774749756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,16384,0.02476799984773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,7168,0.1729724407196045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,512,0.014745778507656522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,32,0.005534222142563925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,1536,0.016531555189026725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,65536,0.6253146595425075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,7168,0.030173334810468886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,7168,0.014364444547229342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,2048,0.020612445142534044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,3072,0.007378666765160031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,512,0.03610310951868693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,12288,0.005222222043408289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,16384,0.03558222121662564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,12288,0.029746668206320867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,12288,0.24340976609124076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,128,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,8192,0.008620444271299574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,65536,0.5026719835069444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,12288,0.019208888212839764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,3072,0.011344888971911537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,8192,0.018188445104493033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,12288,0.06702222426732381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,3072,0.007975999679830339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,4096,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,1024,0.01976711054642995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,12288,0.037350220812691584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,1024,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,5120,0.009646221995353699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,1536,0.005943999936183293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,65536,0.13342132833268908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,768,0.004847110973464118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,1536,0.020232889387342665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,768,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,128,0.0031386667655573953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,5120,0.013277333643701343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,2048,0.014430221584108142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,6144,0.03479111194610596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,6144,0.009299555586444007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,512,0.0024968888610601425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,256,0.015486222174432544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,8192,0.02597155504756504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,3584,0.13065955373975965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,64,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,32,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,1024,0.009544889132181803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,64,0.0037164443896876443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,16384,0.021199999584092036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,3072,0.007269333634111617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,256,0.015107555521859063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,5120,0.008323555191357931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,5120,0.00833155545923445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,16384,0.029143999020258587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,512,0.012784889174832238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,1024,0.0038000001675552795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,8192,0.013442666994200813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,65536,0.16471821731991237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,2560,0.006933333145247565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,2048,0.018534221582942538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,5120,0.009999111294746399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,64,0.005323555734422471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,10240,0.017864000466134813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,768,0.01925244430700938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,3072,0.004230222354332606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,1536,0.020621332857343886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,4096,0.008783111141787635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,12288,0.015431111057599386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,64,0.008738666772842407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,65536,0.06519289149178399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,12288,0.5403573248121474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,3072,0.016196444630622864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,128,0.018238221605618794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,2048,0.027328888575236004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,12288,0.03422310948371887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,7168,0.39974400732252335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,256,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,1024,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,2048,0.006983111302057902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,7168,0.010314666562610203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,6144,0.06767466995451185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,512,0.0111191107167138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,10240,0.012727999852763282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,768,0.004960000101062986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,8192,0.04103555613093906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,3072,0.0191920002301534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,5120,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,2560,0.007264888948864407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,1024,0.005614222337802251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,2048,0.004389333228270213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,256,0.004922666483455234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,128,0.0068693334857622785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,768,0.02516799999607934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,2048,0.004534222185611725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,4096,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,64,0.0021706666383478376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,5120,0.009645333720578088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,1536,0.055564443270365395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,7168,0.08771289057201809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,7168,0.010009778042634329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,2048,0.014160000615649752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,3072,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,2048,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,32,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,8192,0.011725333001878528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,16384,0.03418933351834615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,32,0.005128889034191768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,512,0.005884444548024072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,16384,0.02109866672092014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,3584,0.028956443071365356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,256,0.012800888882742988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,16384,0.02166844407717387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,3584,0.03269866771168179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,7168,0.01036000003417333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,512,0.05244177911016676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,12288,0.021220443977249995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,2048,0.029694222741656836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,12288,0.13364356093936497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,4096,0.007616000043021307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,3072,0.019673777951134574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,10240,0.1425066656536526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,2560,0.01910666624704997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,2048,0.024880889389250014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,2048,0.02198666665289137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,1024,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,10240,0.015480889214409722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,6144,0.021560000048743352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,65536,0.07792177465226915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,10240,0.05637688769234551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,7168,0.01462311049302419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,7168,0.028375112348132666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,7168,0.08046222395367093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,3584,0.019913777709007263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,3584,0.016920889417330425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,7168,0.02977777851952447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,1024,0.007670222057236566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,1024,0.006413333531883027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,768,0.029819554752773706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,512,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,1024,0.005619555711746216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,32,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,3072,0.05470755696296692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,2048,0.014953777194023132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,8192,0.03957688808441162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,2560,0.01995377739270528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,2048,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,8192,0.03919733232922024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,12288,0.015815110670195687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,2048,0.004518222063779831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,8192,0.019127999742825825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,768,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,64,0.00253866674999396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,2048,0.012409778104888068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,512,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,3072,0.014489778214030795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,12288,0.05620355738533867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,32,0.0025253333151340485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,256,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,12288,0.029281778468026057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,65536,0.10562044382095337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,8192,0.045604444212383695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,5120,0.029631111356947158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,4096,0.007662222617202335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,65536,0.057095110416412354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,3584,0.008295111358165741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,768,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,3584,0.013459555804729462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,1024,0.014460444450378418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,128,0.013728000223636627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,4096,0.034394668208228216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,1536,0.013834666874673633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,4096,0.015844444433848064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,12288,0.013937777943081327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,4096,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,6144,0.5798737737867568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,768,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,7168,0.03127288818359375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,1536,0.01221244451072481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,10240,0.020586666133668687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,2048,0.010103999740547603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,768,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,768,0.018239999810854595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,3584,0.019488889310095046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,8192,0.010659555594126383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,3584,0.024031110935741003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,1024,0.017494221528371174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,5120,0.14023021856943765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,256,0.005664000080691443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,1536,0.007253333098358578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,5120,0.008645333349704742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,10240,0.038055999411476985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,8192,0.1496302286783854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,5120,0.012544000314341651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,4096,0.008977777428097194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,128,0.00702311098575592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,10240,0.01204355557759603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,4096,0.43637956513298887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,2560,0.0053591111467944244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,6144,0.015761777758598328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,10240,0.44440534379747176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,3584,0.025495999389224585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,65536,0.09934933318032159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,16384,0.15957244237263998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,768,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,128,0.014468444718254937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,3584,0.011385777758227454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,3072,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,64,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,1536,0.013061333033773633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,128,0.01480888823668162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,16384,0.026605332891146343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,8192,0.052417778306537204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,65536,0.048892445034450956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,3584,0.010006222460005019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,768,0.016694222887357075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,6144,0.009032888544930352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,1024,0.009287111461162567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,2560,0.008294222255547842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,65536,0.40683910581800675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,768,0.003544888976547453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,3072,0.008784888519181145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,256,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,7168,0.010490666660997601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,7168,0.763204468621148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,4096,0.01291911138428582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,65536,0.2426684432559543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,1536,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,65536,0.11402489079369439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,7168,0.009707555174827576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,512,0.005647999958859549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,7168,0.025599999560250178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,1536,0.013267555170589022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,8192,0.008368000388145447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,12288,0.014154665999942355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,64,0.002485333424475458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,4096,0.01605955594115787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,65536,0.07674044370651245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,768,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,512,0.0052266667286554975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,4096,0.010128888818952773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,4096,0.025603555970721777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,1536,0.05762400229771932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,256,0.006035555568006303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,1024,0.01309511148267322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,768,0.017872000734011333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,512,0.00254133323000537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,512,0.013752889302041797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,64,0.01475911173555586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,10240,0.016415110892719693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,3072,0.007301333049933116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,2048,0.007263999846246507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,5120,0.014698666003015308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,768,0.002842666581273079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,6144,0.10520266824298435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,2048,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,3072,0.008329778081840938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,1024,0.00870311094654931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,2560,0.0207315550910102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,128,0.02230311102337307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,8192,0.04193777839342753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,16384,0.04797244403097364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,3072,0.031649778286616005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,6144,0.018788443671332467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,8192,0.04970488945643107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,12288,0.016396444704797532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,4096,0.08826044532987808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,512,0.031368000639809504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,768,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,16384,0.291211552090115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,1536,0.008619555168681675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,7168,0.02242044442229801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,3584,0.008319999608728621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,768,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,2560,0.014493332968817817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,32,0.001839111041691568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,12288,0.07377866903940837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,256,0.002190222135848469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,2560,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,4096,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,1024,0.008405333591832055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,5120,0.1017013324631585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,2048,0.010039110978444418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,768,0.005670222143332164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,12288,0.013821333646774292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,6144,0.019348444210158456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,3584,0.03994399971432156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,768,0.018090666996108163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,512,0.005615111026499007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,3584,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,1024,0.007977777885066139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,1536,0.007134222322040134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,2048,0.004952888935804367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,32,0.005208888815508948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,6144,0.012902222573757172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,6144,0.03534577621353997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,2048,0.020233778489960563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,4096,0.016155555844306946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,768,0.016806221670574613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,6144,0.03119822343190511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,10240,0.04620711008707682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,768,0.002858666703104973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,4096,0.029839111698998347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,1024,0.06155021985371908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,2048,0.011179555621412067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,1024,0.01793244481086731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,4096,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,64,0.002481777841846148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,12288,0.03222577770551046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,2048,0.007937777373525832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,7168,0.008602666358153025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,512,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,3584,0.01198755535814497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,65536,0.027055111196306016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,128,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,256,0.013999111122555204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,5120,0.01640533407529195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,768,0.015878222054905362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,10240,0.034529778692457415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,16384,0.18310488594902885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,8192,0.15066399839189318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,1536,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,65536,0.09468444188435872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,5120,0.021872000561820135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,256,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,10240,0.018048000004556444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,256,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,5120,0.03186755710177951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,8192,0.0214373336897956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,1536,0.013443555268976422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,3584,0.08477599753273858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,4096,0.023361777265866596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,5120,0.0612000028292338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,768,0.0031902222997612427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,6144,0.017161778277821012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,5120,0.02030666669209798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,10240,0.02421422302722931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,512,0.011723555624485016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,768,0.019919999771647982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,2048,0.004580444345871608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,5120,0.01620533400111728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,16384,0.029416001505321924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,4096,0.03291644321547614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,5120,0.004610666798220741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,7168,0.0141004439857271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,2560,0.007176888485749562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,4096,0.05127999848789639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,2048,0.006587555425034628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,10240,0.012076444096035428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,2560,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,1536,0.0299271113342709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,16384,0.023016000787417095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,16384,0.09043200148476495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,1024,0.007607111500369177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,2560,0.22926754421657988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,7168,0.028804444604449805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,64,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,8192,0.06944977574878268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,768,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,128,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,1024,0.014135110709402295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,1024,0.018065777089860704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,7168,0.01959377858373854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,65536,0.09124710824754502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,256,0.023256889647907678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,3584,0.018315555320845712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,1024,0.0032746667663256326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,7168,0.030419554975297716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,3072,0.010778666370444827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,1024,0.015460444821251763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,256,0.0052915554907586836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,1024,0.0075937774446275495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,12288,0.01886755559179518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,8192,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,1536,0.08609155813852946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,768,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,6144,0.016151999433835346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,2048,0.004568888909286923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,3584,0.007814221911960179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,12288,0.013463111387358772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,65536,0.059362669785817467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,64,0.00904799997806549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,8192,0.01850666602452596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,768,0.016913778252071805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,65536,0.10799200005001491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,128,0.002089777754412757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,8192,0.048093332184685596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,512,0.013763555222087435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,5120,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,65536,0.5199271308051215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,768,0.017129777206314933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,1024,0.006215111249023014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,128,0.018527110417683918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,3584,0.06728444496790568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,1536,0.0069395556218094295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,64,0.0042071110672420925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,32,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,1024,0.01716711123784383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,6144,0.019853333632151287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,16384,0.03288888931274414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,1024,0.026776888304286536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,65536,0.11245955361260308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,65536,0.019865777757432725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,3584,0.015504888362354703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,7168,0.012678222523795234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,256,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,768,0.006198222024573221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,256,0.002201777779393726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,6144,0.006638222270541721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,2560,0.008071111308203803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,64,0.005612444546487596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,512,0.011749332977665795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,32,0.007984889050324758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,256,0.005412444472312927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,256,0.007719111111429002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,3584,0.007393777370452881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,65536,0.011736888852384357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,128,0.02160800000031789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,768,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,2048,0.024292444189389546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,6144,0.06781866815355089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,2048,0.007263999846246507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,256,0.0028764444092909494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,1024,0.01276177747382058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,65536,0.05712178018358019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,7168,0.010445333189434476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,2048,0.040031999349594116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,7168,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,32,0.0028151110228565005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,4096,0.016501333978441026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,4096,0.010088889135254754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,1536,0.013752889302041797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,2560,0.007243555453088548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,65536,0.1502764489915636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,512,0.056735999054378934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,1536,0.014112888111008538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,16384,0.17995199892255995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,2560,0.014782221780882942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,12288,0.02050133380624983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,768,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,1536,0.0038248888320393036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,768,0.0028648889727062653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,5120,0.022656889425383672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,8192,0.2802835570441352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,10240,0.3058417903052436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,1536,0.0069155556460221606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,16384,1.7358702553643122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,12288,0.03579555617438422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,128,0.014441778262456259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,4096,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,2560,0.015829333000712924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,768,0.017516444126764934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,1024,0.003379555626047982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,3584,0.007846222155623967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,256,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,12288,0.0236835562520557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,128,0.015438222222858004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,1536,0.006360000206364526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,512,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,10240,0.017788444956143696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,6144,0.014179555906189812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,12288,0.5413733588324653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,2560,0.03326755430963304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,256,0.02922044528855218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,6144,0.009020444419648912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,3072,0.01533066729704539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,256,0.019556444552209642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,5120,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,768,0.0185244447655148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,64,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,3072,0.024784000383483037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,3584,0.04554399847984314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,2560,0.013020444247457715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,256,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,128,0.007307555940416124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,16384,0.4452808698018392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,768,0.0063457778758472866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,6144,0.013787555197874704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,1024,0.00793688909875022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,4096,0.007419555551475949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,768,0.01242844429281023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,128,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,12288,0.00964355551534229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,65536,0.00655644428398874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,2048,0.005977777971161737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,12288,0.019866666860050626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,2560,0.032053334845436945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,3072,0.026104888982243005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,128,0.015846222639083862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,2560,0.06087199846903483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,16384,0.02091022166940901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,6144,0.014757333530320061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,128,0.006621333460013072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,4096,0.01920977731545766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,12288,0.013473778135246701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,512,0.012420444852775998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,768,0.005620444400442972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,3072,0.010311110979980892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,16384,0.021469333105617102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,128,0.0075431110130416016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,5120,0.12580888801150852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,256,0.04247555467817518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,10240,0.012431999875439538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,16384,0.1844151152504815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,32,0.0021395554973019496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,3584,0.008291555775536431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,4096,0.008212444682916006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,32,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,6144,0.010173333187898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,768,0.01647111111217075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,256,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,4096,0.008449777960777283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,512,0.018904000520706177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,16384,0.03264888789918687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,3072,0.0069884442620807225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,5120,0.004538666870858935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,512,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,8192,0.020763556162516277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,64,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,1024,0.01107911103301578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,3584,0.011633777783976661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,5120,0.038458665211995445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,7168,0.040179553959104754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,7168,0.023756444454193115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,256,0.025080889463424683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,2048,0.015465777781274585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,512,0.0052791109515561005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,16384,0.03639377819167243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,256,0.013425778183672162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,3072,0.07876622014575534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,3584,0.01292799992693795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,64,0.0018337778747081757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,6144,0.03992088966899448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,768,0.015626667274369132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,512,0.0025191110455327565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,4096,0.01459111107720269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,10240,0.013369777964221107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,6144,0.009588444398509132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,32,0.01683644453684489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,256,0.005961777849329843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,6144,0.015511999527613321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,1536,0.01907999979125129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,32,0.001869333287080129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,512,0.01035644445154402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,10240,0.04302044378386604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,8192,0.03128977616628011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,3072,0.024029332730505202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,32,0.00654666663871871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,256,0.01185066666867998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,8192,0.011715555356608497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,2048,0.014441778262456259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,8192,0.004193777839342753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,1536,0.027660444378852844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,64,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,512,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,256,0.025415110919210646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,7168,0.13594755861494276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,12288,0.013428444663683573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,512,0.012440000143316058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,10240,0.025617778301239014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,64,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,2048,0.04056266612476773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,7168,0.01756177842617035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,768,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,1536,0.006314666734801398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,4096,0.013888888888888888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,512,0.011336888704035016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,16384,0.03159822358025445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,32,0.018216000662909616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,65536,0.16851022508409289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,768,0.0196115556690428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,8192,0.19486399491628012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,10240,0.03394488824738397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,256,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,65536,1.871503194173177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,8192,0.011732444167137146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,5120,0.014451555079884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,10240,0.02163644466135237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,2560,0.008999110923873054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,128,0.001839111041691568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,7168,0.004853333450025982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,4096,0.3573777675628662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,7168,0.03393955694304572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,7168,0.020727111233605277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,3072,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,8192,0.05682310793134901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,4096,0.008401778009202745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,12288,0.012672888735930124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,16384,0.0647768908076816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,128,0.015126221709781222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,32,0.0031679999083280563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,768,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,128,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,8192,0.03724622064166599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,4096,0.015448000695970325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,3072,0.008278222547637092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,12288,0.026154667139053345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,3584,0.011693333586057028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,7168,0.010002666877375709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,4096,0.013456000222100152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,65536,0.09575911362965901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,4096,0.14642311467064753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,6144,0.04093599981731839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,16384,0.08134755823347303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,16384,0.1338684426413642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,1536,0.054677334096696645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,65536,0.07940000295639038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,2560,0.07220444414350721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,10240,0.055678221914503306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,1024,0.04765244325002035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,768,0.013769778112570444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,1024,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,1536,0.0042088888585567474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,768,0.00797155582242542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,768,0.002887111157178879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,256,0.012415111064910889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,256,0.004587555511130227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,10240,0.02130844526820713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,2048,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,12288,0.016567111015319824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,64,0.001810666587617662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,4096,0.0359208881855011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,12288,0.2040648857752482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,768,0.00564444437623024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,3584,0.014721777704026965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,128,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,10240,0.01257066677014033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,7168,0.04095110959476895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,3584,0.01144177797767851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,2560,0.017613333132531907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,8192,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,256,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,2048,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,5120,0.01647644407219357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,6144,0.009683555199040307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,10240,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,256,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,128,0.005978666659858491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,2560,0.00868888861603207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,2048,0.01958044370015462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,2560,0.023708444502618577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,64,0.005903111149867375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,2560,0.01683555543422699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,8192,0.024499555428822834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,8192,0.012612444659074148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,1024,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,1536,0.013560000393125745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,7168,0.030212445391549006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,8192,0.030216889248953924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,1536,0.018962666392326355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,3072,0.007867555651399825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,12288,0.025723555021815833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,3072,0.02330133318901062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,6144,0.009358222285906473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,6144,0.020580444071027968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,65536,0.016855110724767048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,128,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,1024,0.011730666789743634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,256,0.0059279998143513995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,1536,0.017960889471901786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,768,0.007271111011505127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,8192,0.10285511281755234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,32,0.0069066666894488865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,256,0.01754488878779941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,2048,0.006160888820886612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,1536,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,512,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,768,0.016487111647923786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,128,0.00628977765639623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,12288,0.030184000730514526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,6144,0.011056000159846412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,3072,0.007798222204049428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,128,0.005528888768619961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,256,0.006908444480763541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,512,0.0069155556460221606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,3584,0.033174223370022245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,4096,0.028304888142479792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,65536,0.12371822198232015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,3584,0.008113777471913232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,12288,0.1957777738571167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,4096,0.030188444587919448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,5120,1024,0.022813333405388728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,6144,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,3584,0.6029431025187174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,4096,0.0321724447939131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,128,0.005224888937340842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,10240,0.033779554896884494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,16384,0.23620711432562935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,3584,0.010963555839326648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,1536,0.015798222687509324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,128,0.00795199970404307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,768,0.07268000311321683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,8192,0.011005333728260465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,3072,0.007329777710967594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,2048,0.004277333203289243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,256,0.01647111111217075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,8192,0.006587555425034628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,12288,0.0163093325164583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,12288,0.040728890233569674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,512,0.00925688859489229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,32,0.001847111102607515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,3072,0.018567999203999836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,65536,0.13513155778249106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,1536,0.004588444613748127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,7168,0.024912888805071514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,8192,0.040992889139387347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,65536,0.5547875298394097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,5120,0.012409778104888068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,1024,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,65536,1.283263948228624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,65536,0.06802044312159221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,32,0.012163555456532372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,1024,0.01964977714750502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,3072,0.007644444704055786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,768,0.023990222149425085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,12288,0.01960711181163788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,7168,0.37353867954678005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,256,0.0024968888610601425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,12288,0.030055999755859375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,10240,0.18907554944356283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,768,0.013749333719412485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,1536,0.015449777245521545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,16384,0.021631111701329548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,16384,0.07958044608434041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,10240,0.016880000631014507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,5120,0.011877333124478659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,65536,0.21707644727494982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,2048,0.011395555403497485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,4096,0.02703199949529436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,7168,0.011383111278216044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,3584,0.05764444669087728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,3584,0.014751111467679342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,8192,0.11628266175587972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,4096,0.01551999979548984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,65536,3584,0.15738044844733343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,768,0.05119911167356703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,512,0.02698844505680932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,65536,0.15355111493004694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,256,0.014092443717850579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,16384,0.034662223524517484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,5120,0.02868622210290697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,12288,0.01682844426896837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,3072,0.007520000139872233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,64,0.007649777664078607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,8192,0.044234666559431285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,768,0.017613333132531907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,5120,0.02420266634888119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,4096,0.02530844509601593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,768,0.007463999920421177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,32,0.0052897776994440295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,2048,0.023881778120994568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,512,0.024340444140964087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,8192,0.007271111011505127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,512,0.002587555597225825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,8192,0.011069333387745751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,256,0.004919111314747068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,12288,0.021171554923057556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,10240,0.014115555418862237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,16384,0.1011235581503974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,32,0.008020444048775567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,2560,0.01907466683122847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,12288,0.0444488889641232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,768,0.017813333206706576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,1536,0.011697777443461947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,3584,0.015479111009173922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,256,0.015513777732849121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,10240,0.5745004547966851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,10240,0.20374755064646402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,1024,0.017872000734011333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,1536,0.005050666630268097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,6144,0.03984977801640829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,16384,0.05142044358783298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,768,0.013782222237851886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,768,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,128,0.014448889427714877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,32,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,5120,0.01757422255145179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,3072,0.013756444056828817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,6144,0.02162577708562215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,256,0.005621333503060871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,16384,0.03218133250872294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,2048,0.010009778042634329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,8192,0.01718311177359687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,1536,0.006435555716355641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,7168,0.020710221595234342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,2048,0.13095200061798096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,512,0.010255110760529837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,256,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,2560,0.01372977760103014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,128,0.00600622221827507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,512,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,6144,0.010657777388890585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,7168,0.010640000303586325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,2560,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,12288,0.028390222125583228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,128,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,3072,0.032871110571755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,2048,0.006959999601046245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,6144,0.30641333262125653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,2048,0.005526222288608551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,128,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,32,0.007342221836249034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,128,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,768,0.01796266602145301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,4096,0.007663999994595845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,5120,0.033013333876927696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,16384,0.02071822186311086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,2560,0.008960000342792934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,128,0.014465777410401238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,1024,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,3072,0.03142311175664266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,2560,0.015496888094478183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,7168,0.004861333303981357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,8192,0.015992000699043274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,7168,0.009952888720565373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,256,0.002208000048995018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,3072,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,3072,0.04395022326045566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,128,0.002116444417172008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,32,0.001814222170246972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,5120,0.020908445119857788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,768,0.007141333487298753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,64,0.002132444539003902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,512,0.05667110946443346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,256,0.016893333858913846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,1536,0.06797244151433308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,256,0.005910222315125995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,32,0.002834666727317704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,3072,0.018876444962289598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,1024,0.09384266535441081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,4096,0.014746665954589844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,8192,0.010659555594126383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,65536,0.17473155922359893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,128,0.00865955568022198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,32,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,768,0.002809777855873108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,256,0.0055440002017551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,7168,0.00977244476477305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,768,0.023657777243190344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,5120,0.033575998412238225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,10240,0.03182044294145372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,1536,0.006659555352396435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,32,0.0025315555847353404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,2560,0.013502221968438892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,8192,0.3026871151394314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,4096,0.014193778236707052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,65536,0.06416977776421441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,32,0.014078223043017916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,7168,0.010371555884679159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,65536,0.47714842690361875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,1536,0.004048000193304486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,2048,0.04084533452987671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,256,0.016172443826993305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,6144,0.01090844472249349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,4096,0.03755111164516873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,5120,0.027134221461084154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,5120,0.0957217746310764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,65536,3.201779471503364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,65536,0.18077333768208823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,2560,0.010667555862002902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,1536,0.009691555466916826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,2560,0.013840888937314352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,1024,0.05091822147369385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,4096,0.004910222358173794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,10240,0.036333332459131874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,8192,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,128,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,3072,0.016131555040677387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,512,0.013478221992651621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,128,0.00509066672788726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,2048,0.015111999379263984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,768,0.016755556066830952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,1536,0.026183111800087824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,128,0.011027555498811932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,32,0.009020444419648912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,3072,0.01612622208065457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,768,0.02586577832698822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,2048,0.010066666536860997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,256,0.005220444252093633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,256,0.00591111100382275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,12288,0.019493333167499967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,768,0.017877333694034152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,1024,0.004465777840879229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,3584,0.00489155575633049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,65536,0.7614657613966199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,16384,1.6613875495062933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,16384,0.020593777298927307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,128,0.0028213332924577924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,6144,0.04566399918662178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,4096,0.008642666869693333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,12288,0.059742218918270536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,8192,0.015319110618697273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,1536,0.0303164455625746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,1536,0.0038568890757030914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,12288,0.03957510987917582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,64,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,6144,0.014256889621416727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,1024,0.01646133263905843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,1536,0.012390221986505719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,2560,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,1024,0.027808000644048054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,768,0.0041911109454101985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,1536,0.006587555425034628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,2560,0.0068693334857622785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,64,0.006170666466156642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,32,0.0074240002367231585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,65536,0.08201244142320421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,4096,0.011364444262451597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,8192,0.008960888617568547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,3584,0.01648533344268799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,512,0.006218666831652324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,1024,0.009691555466916826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,16384,0.02345688806639777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,256,0.00832088871134652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,10240,0.00977244476477305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,256,0.01556622154182858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,8192,0.013376000026861826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,6144,0.0655848913722568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,3584,0.007956444389290279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,1024,0.015118221441904703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,2560,0.07876266373528375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,1536,0.014412444498803882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,3072,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,8192,0.010506666368908353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,7168,0.006937777830494775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,7168,0.014444443914625378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,10240,0.042211555772357516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,65536,0.24548710717095268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,128,0.007974222302436829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,512,0.016419554750124615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,64,0.0018826667219400406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,12288,0.013447999954223633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,64,0.007301333049933116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,6144,0.031211555004119873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,2048,0.007296000089910295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,512,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,5120,0.017495999733606975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,8192,0.021586666504542034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,3584,0.018571555614471436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,8192,0.02332088847955068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,6144,0.018541332748201158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,16384,0.16716267002953422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,7168,0.021798221601380244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,5120,0.06002044677734375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,128,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,4096,0.008014221986134848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,2560,0.007631999750932057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,1536,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,65536,2.125354766845703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,8192,0.05095110999213325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,16384,0.018368888232443068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,6144,0.017895999881956313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,2048,0.03248622351222568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,3072,0.01344622257683012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,12288,0.008638222184446123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,5120,0.008966222405433655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,8192,0.014487110906177096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,512,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,3072,0.03206488821241591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,2560,0.03313422203063965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,7168,0.1972151067521837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,12288,0.04772177669737074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,8192,0.5487466918097602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,256,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,65536,0.05693333016501533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,128,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,256,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,8192,0.014104000396198697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,65536,0.0889822244644165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,2048,0.014551111393504672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,128,0.0068195557428730865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,12288,0.01754844519827101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,1024,0.007876444194051955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,128,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,32,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,512,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,8192,0.23342577616373697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,4096,0.010904889139864178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,128,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,1024,0.004914666629499859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,768,0.0070266665683852295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,8192,0.022138666775491502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,3584,0.023947555157873366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,512,0.0032462223122517266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,16384,0.19618577427334258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,5120,0.010303999814722273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,768,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,12288,0.05488533443874783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,3584,0.014125333891974555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,2560,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,12288,0.034435556994544134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,8192,0.02130311065249973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,64,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,128,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,10240,0.011983999775515662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,32,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,65536,0.09342755873998006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,10240,0.034769776794645525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,3072,0.0069226668112807805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,3584,0.018820444742838543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,256,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,512,0.09012799792819554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,64,0.011766222616036734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,8192,0.031473779016070895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,256,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,2048,0.004562666846646203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,2048,0.022259554929203455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,10240,0.06577955351935492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,512,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,512,0.021581333544519212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,2560,0.01735822194152408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,4096,0.0079297779334916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,7168,0.04482311010360718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,10240,0.03557866811752319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,3584,0.020998222960366145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,768,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,2048,0.014140443669425117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,128,0.012376888758606382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,256,0.0028968888024489083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,10240,0.012073777616024017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,7168,0.06254311402638753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,7168,0.021876444419225056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,16384,0.0889128910170661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,2048,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,3584,0.008065777520338694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,1024,0.007978666987684038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,4096,0.008265777594513362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,256,0.01862133377128177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,5120,0.009176000124878353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,1024,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,128,0.016734222571055096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,128,0.01203644441233741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,3584,0.009681777821646797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,7168,0.009986666341622671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,128,0.009480888644854227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,32,0.0058755555914507965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,6144,0.0335440006521013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,12288,0.022457778453826904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,256,0.008270222279760573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,7168,0.009309333231714036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,256,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,2048,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,256,0.002508444504605399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,128,0.01958755486541324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,2560,0.023119999302758112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,2048,0.008329778081840938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,8192,0.010429333481523724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,16384,0.021031111478805542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,32,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,4096,0.007605333295133378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,1536,0.005997333261701796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,1024,0.02313866714636485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,16384,0.01034044474363327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,5120,0.011976000335481433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,5120,0.027549333042568628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,7168,0.19744000169965956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,128,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,6144,0.009313777916961247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,5120,0.018725333942307364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,64,0.001820444439848264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,65536,0.1482293340894911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,3072,0.013227555486891003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,4096,0.018404444058736164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,1024,0.018179555733998615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,10240,0.054903109868367515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,3584,0.01127377814716763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,768,0.0057751110030545135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,8192,0.020223110914230347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,5120,0.004851555658711327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,1024,0.01443199978934394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,1536,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,64,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,3584,0.10612089104122585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,16384,0.016520000166363187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,8192,0.033667554457982383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,1024,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,16384,0.03936800029542711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,256,0.005959110955397288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,8192,0.03282666537496779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,65536,0.18194399939643013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,16384,0.4575893084208171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,512,0.03086577852567037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,128,0.0018942221585247253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,3584,0.007982222570313348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,8192,0.0505964457988739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,8192,0.014894222219785055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,12288,0.020233778489960563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,1536,0.0236835562520557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,10240,0.027035555905765955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,12288,0.041167110204696655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,512,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,16384,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,2048,0.011172444456153445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,2048,0.07064177592595418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,512,0.02188088827662998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,32,0.0020862221717834473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,6144,0.008819555242856344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,1024,0.015267555912335714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,10240,0.012028444144460889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,256,0.014753777119848462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,12288,0.32101334465874565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,1536,0.11689421865675186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,128,0.015785778562227886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,768,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,6144,0.09030133485794067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,768,0.017497777938842773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,16384,0.006028444402747684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,128,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,10240,0.007990222010347579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,512,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,7168,0.017335999343130324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,65536,0.02808088726467556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,16384,0.41495111253526473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,16384,0.0640231106016371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,256,0.013248888982666863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,32,0.006986666884687211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,65536,0.07403644588258532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,10240,0.01573777695496877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,2560,0.008444444172912175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,5120,0.008612444831265343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,3584,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,6144,0.01721155477894677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,3072,0.015405333704418607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,128,0.0018977777411540349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,1536,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,2048,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,512,0.005614222337802251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,6144,0.05973955657747057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,768,0.005949333310127258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,4096,0.004571555389298333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,256,0.0024826667375034755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,3584,0.03872355487611558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,768,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,4096,0.010827555424637265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,2560,0.006559111177921295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,2048,0.012439111040698158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,1536,0.003489777859714296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,65536,1.510032018025716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,6144,0.027941332923041448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,3072,0.013171555267439948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,2048,0.007885333564546373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,12288,0.030839999516805012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,6144,0.027436445156733196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,32,0.004281777888536453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,1536,0.003528888854715559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,65536,0.07349066601859199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,8192,0.03277511066860623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,2560,0.044871111710866295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,16384,0.053132444620132446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,64,0.001826666709449556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,8192,0.017214222086800467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,3584,0.011025778121418424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,128,0.0196115556690428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,128,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,7168,0.14648355378044978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,1024,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,3584,0.005323555734422471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,2048,0.02438844409253862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,768,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,5120,0.008998221821255153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,5120,0.02197066611713833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,768,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,5120,0.013352000051074557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,768,0.013394667042626275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,256,0.005530666559934616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,2048,0.006597333484225803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,128,0.013420444395807056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,12288,0.018248000078731112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,3072,0.056052446365356445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,5120,0.008639111287064022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,6144,0.01239466667175293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,3072,0.12441866927676731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,4096,0.023220444718996685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,2560,0.015132443772421943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,768,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,10240,0.9106168746948242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,8192,0.09762933519151475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,3072,0.008268444074524773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,1024,0.017836444907718234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,5120,0.03189422355757819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,10240,0.012077333198653327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,128,0.008642666869693333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,12288,0.018586667047606576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,64,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,8192,0.027041777968406677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,3072,0.026737777723206416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,12288,0.049235555860731334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,3072,0.017183999220530193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,5120,0.01016355554262797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,4096,0.018228444788191054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,3584,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,32,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,2048,0.7829573419358996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,64,0.006948444578382704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,12288,0.01740800009833442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,1536,0.007989333735571967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,8192,0.015592889653311836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,768,0.003528000166018804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,768,0.007329777710967594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,16384,0.04447466797298855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,16384,0.04470577836036682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,512,0.011069333387745751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,3584,0.025423111187087163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,256,0.008623999853928884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,5120,0.027074666486846075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,8192,0.011332444846630096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,1536,0.006640888750553131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,10240,0.019478221734364826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,2048,0.012720000412729053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,128,0.002064888882968161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,512,0.02029244436158074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,768,0.003543111185232798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,2560,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,65536,1.0691199832492404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,2560,0.16724267270829943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,8192,0.019928889142142404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,64,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,4096,0.014892444014549255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,512,0.0052800000541739995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,256,0.015294222368134392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,2560,0.006694222076071634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,65536,0.8668577935960559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,2048,0.02478666603565216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,3072,0.009001777403884465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,32,0.0021280000607172647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,6144,0.018889778190188937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,512,0.01443288889196184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,16384,0.1026631130112542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,2560,0.004907555464241239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,128,0.0031928889867332247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,16384,0.020074667202101815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,12288,0.2963457902272542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,5120,0.01720622181892395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,3072,0.02995022137959798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,2560,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,32,0.0052044445441828836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,5120,0.011977777712874942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,8192,0.01072177787621816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,3584,0.04847466614511278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,512,0.006051555689838197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,3584,0.07005066341824003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,2560,0.03405155407057868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,10240,0.019706666469573975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,2048,0.02678311036692725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,768,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,12288,1.2131421830919054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,16384,0.37827467918395996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,1024,0.005097777893145879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,5120,0.01382844481203291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,128,0.002847111059559716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,512,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,4096,0.0459191103776296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,2560,0.03937600056330363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,16384,0.15226133664449057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,12288,0.02329066726896498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,4096,0.0273964438173506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,12288,0.013774221969975365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,32,0.06679733594258626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,5120,0.01070755554570092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,32,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,2560,0.00758844448460473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,1024,0.019555555449591745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,4096,0.04887110988299052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,768,0.003136000078585413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,2560,0.006888888776302338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,3584,0.01791822248035007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,12288,0.028484443823496502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,3584,0.023224888576401606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,512,0.005240889059172736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,1536,0.023296000228987798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,3072,0.007826666865083907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,5120,0.028241776757770117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,7168,0.08487288819419013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,7168,0.013215999636385175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,32,0.0019013333237833446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,1024,0.007391110890441471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,12288,0.01984355515903897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,2560,0.0070159998204973005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,2560,0.012421333127551608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,7168,0.008622222476535374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,256,0.006559111177921295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,1024,0.006518222391605377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,64,0.0029360000044107437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,32,0.0021297778520319196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,2560,0.050476445092095264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,65536,0.019144000278578866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,3584,0.10524177551269531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,5120,0.008314666648705801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,10240,0.07646933529112074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,2048,0.008676444490750631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,256,0.011946666571829053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,3072,0.010114666488435533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,10240,0.08470400174458821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,2048,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,12288,0.009923555784755284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,128,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,256,0.005208000126812193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,128,0.0031600000543726813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,65536,0.0686159994867113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,1024,0.012077333198653327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,32,0.00281866660548581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,4096,0.012960000170601739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,2560,0.005229333208666907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,3072,0.007965332931942409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,512,0.0083155557513237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,2560,0.007254222200976477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,6144,0.02260355485810174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,3584,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,32,5120,0.004563555535342959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,64,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,1024,0.00490044429898262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,4096,0.019913777709007263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,2560,0.012741333080662621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,1024,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,5120,0.1962826649347941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,8192,0.02609599961174859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,10240,0.010068444742096795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,256,0.002550222186578645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,65536,0.07052444749408297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,6144,0.013780444032616086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,5120,0.008608889248636033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,10240,0.13615555233425564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,7168,0.019676445258988273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,3584,0.018176888426144917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,7168,0.009650666680600908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,8192,1.5518915388319228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,3584,0.05367999937799242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,1536,0.01552977826860216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,3072,0.005271111097600725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,512,0.00453244439429707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,1536,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,4096,0.017495110630989075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,16384,0.34121065669589573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,6144,0.03000266684426202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,64,0.002448888909485605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,6144,0.030572444200515747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,1024,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,64,0.009134222235944536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,1024,0.022637334134843614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,16384,0.04042133357789781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,2048,0.022118222382333543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,6144,0.024181332853105333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,512,0.022456889351209004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,2048,0.006626666833957036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,3072,0.17300355434417725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,1536,0.01969688965214623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,32,0.0021546667234765156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,768,0.013177778157922955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,3584,0.019552888141738046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,2560,0.021172444025675457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,7168,0.005534222142563925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,8192,0.01945955554644267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,10240,0.022022222479184467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,10240,0.5020444658067491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,65536,0.5163804690043131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,10240,0.40276087654961484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,7168,0.022835556003782485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,2048,0.006575111299753189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,3072,0.007255111303594377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,3072,0.026161778304311965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,1024,0.018581334087583754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,8192,0.023078221413824294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,8192,0.025045333637131587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,768,0.007610666255156199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,16384,0.02272622287273407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,8192,0.10580711232291327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,12288,0.029858665333853826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,10240,0.020469332734743755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,512,0.0032097777972618737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,512,0.005401777724424998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,256,0.00795555528667238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,512,0.014463999205165438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,7168,0.01536355581548479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,12288,0.06865155696868896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,2560,0.015320888823933072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,3584,0.021014221840434607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,8192,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,7168,0.017779555585649278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,64,0.005184000151024925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,1536,0.015307555596033732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,7168,0.012786666552225748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,128,0.05573155482610067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,6144,0.03388088941574097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,3072,0.01627377834584978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,4096,0.026480888326962788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,3584,0.01973511113060845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,128,0.00555644432703654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,64,0.005424889011515512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,512,0.002480888946188821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,768,0.0079297779334916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,512,0.016855110724767048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,6144,0.009374221993817223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,7168,0.01721600029203627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,128,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,12288,0.01868266695075565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,5120,0.02940355406867133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,512,0.0029066666546795103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,7168,0.026878222823143005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,16384,0.016460445192125108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,6144,0.01886755559179518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,12288,0.04930488930808174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,1536,0.01816444430086348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,65536,0.11214933130476211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,3584,0.007893333004580604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,2048,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,65536,0.30265866385565865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,3072,0.008955555657545725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,12288,0.04031822085380554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,7168,0.011378666592968835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,16384,0.06362044148974948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,3072,0.00793599999613232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,2560,0.03465955456097921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,65536,0.12649867269727919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,7168,0.012318222059143914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,512,0.017179555363125272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,12288,0.10604888863033718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,3072,0.010499555203649731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,32,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,3072,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,1536,0.019664888580640156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,12288,0.012123555772834353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,512,0.0031964443624019623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,256,0.0028373334142896864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,32,0.09979910982979669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,128,0.015138667490747241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,1024,0.009628444082207149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,4096,0.00830844458606508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,2048,0.008995555341243744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,4096,0.004551110996140374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,3584,0.01889333294497596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,7168,0.017882666654056974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,1024,0.006719999843173557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,2560,0.006578666468461354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,256,0.007997333175606197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,2048,0.008917333351241218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,512,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,8192,0.020612445142534044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,1536,0.00963288876745436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,16384,0.022083555658658344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,3584,0.006296000132958095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,32,0.002817777709828483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,1024,0.004857777721352047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,2560,0.02860977914598253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,768,0.0036515556275844574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,1536,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,5120,0.01795288920402527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,64,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,256,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,256,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,16384,0.041639112763934664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,1024,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,7168,0.12887822257147893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,6144,0.033915556139416166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,8192,0.01721600029203627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,2048,0.0069751110341813825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,1024,0.008619555168681675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,7168,0.02586488922437032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,16384,0.11429866154988606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,256,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,768,0.017870222528775532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,1536,0.009674666656388177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,512,0.01204711116022534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,2048,0.05177244544029236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,3072,0.029486222399605647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,5120,0.01385244478782018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,64,0.00797333319981893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,4096,0.018939556346999276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,128,0.015111999379263984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,768,0.002883555574549569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,128,0.007639110916190677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,3072,1.0795129140218098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,5120,0.013293333351612091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,2048,0.034725334909227155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,8192,0.09027555253770615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,128,0.05423644516203138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,3072,0.01590666671593984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,12288,0.03591288791762458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,128,0.0018648888087934917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,256,0.0021928888228204516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,16384,0.06479911009470622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,16384,0.03596533338228861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,256,0.037314666642083064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,512,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,256,0.016501333978441026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,65536,0.12625332673390707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,6144,0.12200889322492812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,12288,0.031222220924165513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,2560,0.05573599868350559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,4096,0.03219199842876858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,5120,0.016535111599498324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,2560,0.012651555240154266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,768,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,6144,0.03727733426623874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,256,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,8192,0.01310755560795466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,10240,0.013410666750537025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,768,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,5120,0.00924177798959944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,256,0.0024622221373849446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,7168,0.02201244400607215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,3584,0.01107911103301578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,3584,0.004608889006906086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,5120,0.008998221821255153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,5120,0.013067555924256643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,2560,0.006993778049945831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,65536,0.10118222236633301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,3072,0.02463111115826501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,768,0.05524533324771457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,256,0.01513777838812934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,2048,0.00425511101881663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,1024,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,1024,0.028731554746627808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,16384,0.036455111371146307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,3584,0.033555554019080266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,256,0.030850665436850652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,2560,0.0075635554061995606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,2560,0.011483555866612328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,1024,0.007740444607204861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,128,0.0028435554769304064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,256,0.01514755520555708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,4096,0.05707999732759264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,7168,0.019837333096398246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,1536,0.01963555481698778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,8192,0.010939555863539377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,128,0.013056000073750814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,10240,0.012058667010731168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,256,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,3584,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,512,0.01960266629854838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,2048,0.00664533343580034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,6144,0.04044622182846069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,2560,0.00740444411834081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,5120,0.016144000821643405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,4096,0.10390844610002305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,2048,0.004605333424276776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,256,0.015116444892353482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,16384,0.028491556644439697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,7168,0.019913777709007263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,3584,0.027395556370417278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,768,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,4096,0.016695110334290397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,10240,0.025480000509156123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,7168,0.035306665632459856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,10240,0.0266942216290368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,5120,0.013769778112570444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,512,0.002481777841846148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,6144,0.012737777498033313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,512,0.0035297779573334586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,1536,0.007670222057236566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,64,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,8192,0.03689600030581156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,10240,0.05183022220929464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,1536,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,16384,0.005611555443869696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,2560,0.016546666622161865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,10240,0.18471999963124594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,6144,0.02535022298494975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,1536,0.0034844444857703317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,2048,0.008671110702885522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,16384,0.042928887738121875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,2048,0.004531555705600315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,1024,0.004557333472702238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,16384,0.02516533268822564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,3072,0.015104888214005364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,8192,0.015805333852767944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,3072,0.015101333459218344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,1536,0.014261333478821648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,256,0.004880889008442561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,4096,0.38955733511183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,16384,0.01554666625128852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,64,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,768,0.01959288948112064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,1536,0.012856889102194043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,12288,0.021055110626750525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,256,0.005948444621430502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,256,0.005935110979610019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,16384,0.05013422171274821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,3584,0.04679110977384779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,2560,0.4224533240000407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,4096,0.006583111153708563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,65536,0.19566222031911215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,768,0.02132533325089349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,12288,0.01771111124091678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,10240,0.025106666816605464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,128,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,7168,0.010318222145239512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,3072,0.026554667287402686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,3584,0.010423111418883005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,768,0.0055440002017551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,2560,0.015298666225539314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,768,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,128,0.014856000741322836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,2048,0.0063457778758472866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,1536,0.01202222208182017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,512,0.014114666316244336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,4096,1.490672853257921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,16384,0.026458667384253606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,10240,0.04030133287111918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,128,0.01165600038237042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,6144,0.09972533252504136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,128,0.0028195555011431375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,65536,1.6496950785319011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,12288,0.05794577466117012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,128,0.008931555681758458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,2560,0.07350489166047838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,512,0.0028195555011431375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,3584,0.008593777815500895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,2560,0.014374223020341663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,1536,0.006899555524190267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,12288,0.07149689065085517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,32,0.002181333386235767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,3584,0.08257955312728882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,256,0.016513778103722464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,16384,0.025388444463411968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,12288,0.025012445118692186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,256,0.010584000084135266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,4096,0.03952622082498338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,4096,0.01131199962562985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,7168,0.012452444268597497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,3584,0.06047022342681885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,3072,0.09685955444971721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,128,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,6144,0.011729777687125735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,7168,512,0.006467555546098285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,1536,0.058373332023620605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,10240,0.02347911066479153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,6144,0.020781333247820537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,2048,0.00666222224632899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,2048,0.008824889030721452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,2048,0.06465688678953382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,64,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,7168,0.0042204442951414315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,5120,0.08721066845787896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,2560,0.012311999996503195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,65536,0.21821867095099556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,3584,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,128,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,2560,0.012713777522246042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,3584,0.014936000108718872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,2560,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,6144,0.013014222184816996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,65536,0.20182489024268258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,10240,0.09473155604468451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,16384,0.24245956208970812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,1024,0.07551378011703491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,512,0.0069848886794514125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,3072,0.007272889216740926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,8192,0.010675555302037133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,32,0.006265777680608962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,5120,0.014510222607188754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,4096,0.011753777662913004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,128,0.014429334137174817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,128,0.014838222000333997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,1024,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,128,0.01962577799956004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,16384,0.0192266669538286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,128,0.011810666984981961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,256,0.015298666225539314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,2560,0.02202933364444309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,3584,0.00741511086622874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,32,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,4096,0.02789688772625393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,7168,0.052143110169304736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,65536,0.06492177645365398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,768,0.002464888824356927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,768,0.04429866539107429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,32,0.01443199978934394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,7168,0.0074675555030504865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,1536,0.025864000121752422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,5120,0.011595555477672152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,12288,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,16384,0.023228444986873206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,512,0.05628089110056559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,1024,0.02437422176202138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,4096,0.008008889026112026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,2048,0.05741955836613973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,65536,0.15099288357628718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,256,0.014102222190962898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,10240,0.012392000191741519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,3072,0.03840355409516229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,256,0.020242666204770405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,5120,0.04007555709944831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,1024,0.016231111354298063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,16384,0.01688888834582435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,6144,0.10689600308736165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,5120,0.29160798920525444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,5120,0.015167999598715039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,2560,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,5120,0.027638221780459087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,128,7168,0.021078222327762183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,2048,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,65536,0.04214400053024292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,1536,0.009707555174827576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,7168,0.01904622217019399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,3072,0.012191111014948951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,10240,0.024342222346199885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,256,0.0058702222175068325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,128,0.0021991110924217436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,7168,0.18263288338979086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,1024,0.007895111209816402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,16384,0.026900443765852187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,10240,0.1749511162439982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,256,0.0059164443777667145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,768,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,7168,0.02176622218555874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,6144,0.5332106484307183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,12288,0.021331555313534204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,12288,0.03219822380277846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,128,0.013758222262064615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,32,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,4096,0.011634666886594562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,1536,0.0036684444381131064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,2048,0.011729777687125735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,512,0.00254844439526399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,128,0.011355555719799466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,3584,0.00490133340160052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,8192,0.011039111349317761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,10240,0.02916533417171902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,10240,0.037476443582110934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,3072,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,16384,0.05045244428846571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,6144,0.011051555474599203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,5120,0.004559111264016894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,6144,0.005883555445406172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,2048,0.007163555257850223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,2048,0.010012444522645738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,2048,0.006568000134494569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,65536,0.0816755559709337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,2560,0.007080000307824876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,64,0.00553688903649648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,16384,0.023545778459972803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,8192,0.004947555561860402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,10240,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,768,0.01614488826857673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,1536,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,64,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,6144,0.055035554700427584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,12288,0.025783111651738484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,5120,0.02100533412562476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,8192,0.01350044459104538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,4096,0.007890666524569193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,12288,0.036441776487562395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,2560,0.2222773234049479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,12288,0.035262223747041486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,8192,0.01387466655837165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,6144,0.01423733267519209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,3584,0.01683466633160909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,6144,0.22844444380866158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,1536,0.006305777778228124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,4096,0.02700622214211358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,2560,0.01700266698996226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,8192,0.048968887991375394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,1536,0.00906311141120063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,256,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,4096,0.005884444548024072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,3072,0.00572622236278322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,32,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,2560,0.008033778104517195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,5120,0.011683555940786997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,5120,0.017993777990341187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,10240,0.025467554728190105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,32,0.002521777732504739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,128,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,2048,0.014454222387737699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,10240,0.05106133222579956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,7168,0.028891556792789038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,16384,0.01036622209681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,12288,0.04785511228773329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,1536,0.003571555432346132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,2048,0.023277334041065637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,16384,0.02437244521247016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,16384,0.06008711126115587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,256,0.007281777759393056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,3072,0.01754133403301239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,512,0.01646933290693495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,65536,0.0664657751719157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,7168,0.009635555247465769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,1024,0.008000888758235509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,5120,0.01773955590195126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,2560,0.011261333194043903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,5120,0.012852444416946836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,1024,0.01609422266483307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,6144,0.03206400076548258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,256,0.039212445418039955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,1024,0.015506666567590503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,768,0.00721688899728987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,12288,0.026125333375400964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,7168,0.01255022237698237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,6144,0.018195556269751657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,4096,0.0147386673423979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,256,0.015492444237073263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,3072,0.0391537778907352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,5120,0.026512000295850966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,32,0.010379555324713389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,512,0.012404444317022959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,2048,0.013087111214796701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,3072,0.01001422190003925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,512,0.009881777895821465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,2560,0.012072000238630505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,2048,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,16384,0.009213333328564962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,256,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,8192,0.1763066715664334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,7168,0.020985777179400127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,1536,0.006215111249023014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,12288,0.018608889646000333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,256,0.013789333403110504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,6144,0.040919111834632024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,256,0.01510933372709486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,512,0.011210666762457954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,1024,0.009579555855857002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,512,0.0024844445288181305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,8192,0.19990400473276773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,6144,0.05398311217625936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,65536,0.10880800088246663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,128,0.005403555515739653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,7168,0.026349332597520616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,12288,0.052352888716591724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,1536,0.024336889386177063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,3072,0.016474665866957772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,1024,0.039798223310046725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,2048,0.024170666933059692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,3584,0.008218666745556725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,512,0.005716444303592046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,2048,0.04708533485730489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,12288,0.013746666411558786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,12288,0.013394667042626275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,2560,0.01237511138121287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,65536,0.15368710623847112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,8192,0.012457778056462606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,5120,0.00925066653225157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,512,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,768,0.026173333326975506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,3584,0.028110222684012517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,1536,0.13506578074561224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,3584,0.3013937738206652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,1024,0.006088888893524806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,3584,0.005610666755172942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,10240,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,128,0.022329777479171753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,64,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,128,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,16384,0.02131822208563487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,3072,0.004231111043029361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,512,0.017237333787812125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,1536,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,7168,0.037252446015675865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,768,0.0032097777972618737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,4096,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,65536,0.6036551263597276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,10240,0.26011112001207143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,32,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,1024,0.004519999855094486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,16384,0.08064266708162096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,1536,0.0038524443904558816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,12288,0.013431111143694984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,7168,0.004912888838185204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,768,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,8192,0.021007999777793884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,8192,0.09077155590057373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,2560,0.03258666727277968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,5120,0.015458666616015963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,1024,0.006249777972698212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,64,0.001820444439848264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,8192,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,768,0.0031591111587153543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,32,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,768,0.0030533334033356774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,1024,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,3072,0.007281777759393056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,768,0.03571733170085483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,12288,0.009387555221716562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,768,0.017688888642523024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,6144,0.014099554883109199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,6144,0.017667555146747164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,2560,0.02260711126857334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,3072,0.012415111064910889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,256,0.0021742222209771476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,3584,0.007599111232492659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,12288,0.01680355601840549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,1536,0.007292444507280986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,65536,0.0656906697485182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,16384,0.02549155553181966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,3072,0.009952888720565373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,2560,0.012135110795497894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,2048,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,16384,0.6289928754170736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,4096,0.022434666752815247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,65536,0.15525155597262913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,1024,0.015448000695970325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,10240,0.02780088782310486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,1536,0.007986666427718269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,1024,0.003141333245568805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,2048,0.05798400110668606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,1024,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,6144,0.009302222066455418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,4096,0.014631110760900708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,4096,0.007710222568776872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,16384,0.015487111277050443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,512,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,1024,0.01237066669596566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,3072,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,512,0.002485333424475458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,3072,0.0147724449634552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,16384,0.01569422251648373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,3584,0.16255556212531194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,2048,0.00833333366447025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,4096,0.060770663950178355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,512,0.05617066886689928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,1536,0.008709333009190029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,3584,0.027063111464182537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1024,3072,0.015104888214005364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,8192,0.0244835548930698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,8192,0.022816000713242426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,3584,0.014418666561444601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,2048,0.007869333028793335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,128,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,2560,0.01979377700222863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,12288,0.01714666684468587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,65536,0.06425778071085612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,5120,0.01772533357143402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,16384,0.05806755357318454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,32,0.010047111246320937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,6144,0.02106844385464986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,6144,0.018181333939234417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,3584,0.011129777464601727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,512,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,2560,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,5120,0.028577778074476454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,5120,0.02604444490538703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,4096,0.02346666653951009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,768,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,1536,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,256,0.005618666609128316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,8192,0.06944977574878268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,3072,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,3584,0.059715555773841016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,32,0.008228444390826756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,16384,0.02036711076895396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,2048,0.007409777906205919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,4096,0.008112000094519721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,12288,0.017626666360431247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,3072,0.013623110949993134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,256,0.002805333377586471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,7168,0.014130666851997375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,2048,0.08078222142325507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,512,0.05750044186909994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,7168,0.025664000047577754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,2048,0.022807111342748005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,7168,0.010016889207892949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,16384,0.005588444570700328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,3072,0.017889777819315594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,1024,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,6144,0.009653333160612319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,2560,0.006992888947327931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,16384,0.030585779084099665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,8192,0.015446222490734525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,12288,0.0052577778697013855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,5120,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,3584,0.020602666669421725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,10240,5120,0.34748087988959414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,512,0.006909333169460297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,1024,0.007621333003044128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,5120,0.009010666774378883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,768,0.005535999933878581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,10240,8192,0.04218133290608724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,32,0.0052737775776121355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,2560,0.057568887869517006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,3072,0.004566222015354368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,256,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,10240,0.03997600078582764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,10240,0.013784888717863293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,512,0.0148053334818946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,12288,0.016152888536453247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,7168,0.027591110931502447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,16384,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,512,0.005809777726729711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,2048,0.07405689027574328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,1024,0.005774222314357758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,2560,0.029437333345413208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,64,0.0024746666765875286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,10240,0.0388080014122857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,12288,0.06151377492480808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,8192,0.19984889030456543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,3584,0.016879111528396606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,7168,0.011704888608720569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,3072,0.014787554740905762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,2560,0.00832177781396442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,7168,0.021482666333516438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,2048,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,2048,0.00424177779091729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,8192,0.016125332978036668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,2560,0.027235555979940627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,512,0.004179555508825514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,7168,0.02106133268939124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,8192,0.030955556366178725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,16384,0.025677333275477093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,5120,0.04693866769472758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,4096,0.035903112755881414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,4096,0.01035822182893753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,32,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,6144,0.01090133355723487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,1536,0.015490666031837463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,8192,0.033933334880405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,2048,0.025410667061805725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,12288,0.017288888494173687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,12288,0.013175110850069257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,8192,0.010703110860453712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,1536,0.0063662222690052455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,6144,0.0069520001610120135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,3584,0.016509332590632968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,128,0.007306666837798224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,768,0.007259555160999298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,5120,0.016359110673268635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,256,0.006088888893524806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,8192,0.23750933011372885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,5120,0.009768000079525841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,5120,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,2048,0.00676711110605134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,12288,0.05094844433996412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,16384,0.05929511123233371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,16384,0.03788888785574172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,10240,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,12288,0.23144888877868652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,7168,0.020566221740510728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,2048,0.0068986668354935115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,10240,0.06855911016464233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,4096,0.020631111330456205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,1536,0.008094222181373173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,5120,0.017528888252046373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,256,0.002467555511328909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,6144,0.024513777759340074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,128,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,12288,0.013754666679435305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,5120,0.010260444548394945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,7168,0.005439110928111606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,3584,0.01035377797153261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,12288,0.01663733356528812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,12288,0.19942488935258654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,10240,0.209443555937873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,16384,0.03793599870469835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,768,0.01719466679626041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,2048,0.004581333448489507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,8192,0.024586665961477492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,5120,1.0643368826972115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,8192,0.010610666539933948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,12288,0.03262666530079312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,4096,0.007989333735571967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,2048,0.007248000138335758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,512,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,1536,0.004304000073009067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,8192,0.04064977831310696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,64,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,65536,0.2529555426703559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,65536,0.1875244379043579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,4096,0.007657777931955125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,512,0.022143110632896423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,8192,0.157258669535319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,2560,0.01921955578856998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,1024,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,16384,0.022729777627521094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,5120,0.020340444313155282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,3072,0.015807999504937064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,2048,0.01954844428433312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,4096,0.04547911220126682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,128,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,1536,0.00620177760720253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,7168,0.04751288890838623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,12288,0.31690043873257107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,512,0.0069217777086628815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,1024,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,64,0.002480000050531493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,7168,0.03292177783118354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,128,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,64,0.0018684443914228016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,1024,0.00830577810605367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,1536,0.017902221944597032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,65536,0.13705510563320583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,10240,0.03645777702331543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,768,0.005700444595681296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,1024,0.009646221995353699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,1536,0.017860444055663217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,256,0.01511111193233066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,5120,0.016069332758585613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,2048,0.02500266664557987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,65536,0.07778844568464491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,10240,0.02369511127471924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,2048,0.018914666440751817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,8192,0.09692711300320095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,768,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,10240,0.03096888793839349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,512,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,16384,0.03169599837727017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,7168,0.02401155564520094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,10240,0.010321777727868822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,3072,0.016428444120619033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,1024,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,1536,0.006246222390068903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,10240,0.016837333639462788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,3072,0.021518222159809534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,1024,0.008314666648705801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,256,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,8192,0.025776000486479864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,1536,0.006578666468461354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,1536,0.006259555617968242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,256,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,3584,0.007607111500369177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,2560,0.04651733239491781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,6144,0.028772443532943726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,65536,0.18290755483839247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,10240,0.007992000215583378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,1536,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,1024,0.36760711669921875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,256,0.0025048889219760895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,128,0.019713777634832595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,8192,0.04148799843258328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,768,0.007896000312434303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,512,0.21282577514648438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,2560,0.007822222179836696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,8192,0.16486667262183294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,3072,0.00865955568022198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,8192,0.11234577496846516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,8192,0.046295109722349376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,8192,0.011368888947698804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,10240,0.012103111379676394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,768,0.0151573336786694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,6144,0.1740586625205146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,2560,0.02055999967787001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,512,0.006604444649484422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,16384,0.03870044483078851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,3072,0.019739554988013375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,65536,0.18195022477043998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,16384,0.04772444566090902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,1536,0.015202666322390238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,7168,0.0329422222243415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,32,0.014202665951516895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,2048,0.015850666496488783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,2560,0.011027555498811932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,8192,0.020656888683636982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,128,0.014494222071435718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,12288,0.0269128895468182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,2560,0.03600355651643541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,2560,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,1024,0.007611555357774098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,65536,0.07748266723420885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,128,0.005042666776312722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,1024,0.024988444315062627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,256,0.002505777817633417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,10240,0.04443200098143684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,128,0.011352000137170156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,4096,0.011688888900809817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,128,0.0021226666867733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,7168,0.012929778132173749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,768,4096,0.017491555876202054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,6144,0.040800889333089195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,1024,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,8192,0.04064888755480448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,2048,0.004247111164861255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,7168,0.21575021743774414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,7168,0.15562132994333902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,4096,0.011566222541862063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,64,0.005712000032265981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,256,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,7168,0.011733333269755045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,7168,0.14710133605533177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,128,0.014807111687130399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,7168,0.014098667436175875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,16384,0.016524444023768108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,1024,0.007359111474619971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,3584,0.015459555718633862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,3072,0.013032000097963544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,65536,0.016181333197487723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,128,0.015465777781274585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,16384,0.1411528852250841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,1024,0.14139466815524632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,65536,0.07621600230534871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,65536,0.08892799748314752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,5120,0.008829333715968663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,512,0.011044444309340583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,8192,0.030040889978408813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,512,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,8192,0.02994399931695726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,16384,0.04118044508828057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,8192,0.010975110861990186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,65536,0.552694214714898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,1024,0.009748444788985783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,12288,0.03352355625894334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,10240,0.04345422320895725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,3072,0.010303999814722273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,3584,0.02588711182276408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,16384,0.017087111870447796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,10240,0.04547199938032362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,1024,0.00295733329322603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,64,0.0019697778754764134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,256,0.0028817777832349143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,1024,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,7168,0.2491990990108914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,16384,0.23178309864468047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,6144,0.02714577813943227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,512,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,32,0.00244711111817095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,512,0.013784888717863293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,5120,0.04904266529613071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,1024,0.015847999188635085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,65536,1.076541370815701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,1024,0.014085332552591959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,4096,0.0533902214633094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,12288,0.020875554945733812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,2560,0.022486221459176805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,64,0.001855111163523462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,7168,0.025066667132907446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,5120,0.011075555450386472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,2048,0.024784000383483037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,2048,0.03239377670817905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,2560,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,2048,0.009658666948477427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,65536,1.3968097898695204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,512,0.014489778214030795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,5120,0.03391822179158529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,128,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,12288,0.03063911199569702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,256,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,10240,0.04202933443917168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,12288,0.10318844185935126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,65536,0.07505955298741658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,10240,0.013070222404268054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,512,0.006595555692911148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,128,0.0021280000607172647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,7168,0.010050666828950247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,256,0.005525333185990651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,1536,0.0038480001191298165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,12288,0.01793155570824941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,64,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,512,0.014544889330863953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,128,0.006599111275540457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,8192,0.11010755432976617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,1024,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,1536,0.010194666683673859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,4096,0.01646311084429423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,512,0.014885332849290637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,2560,0.014450665977266101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,10240,0.30036889182196724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,1536,0.028664890262815688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,65536,0.05655644337336222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,3584,0.021371554997232225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,2048,0.01484444406297472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,7168,0.36081245210435653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,32,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,10240,0.03822311096721225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,3584,0.016884444488419425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,3584,0.007970666719807519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,32,0.03906222184499105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,8192,0.023103111320071753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,2048,0.017872000734011333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,3584,0.010848888920413123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,6144,0.021838222940762837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,128,0.006196444647179709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,64,0.0021582223061058256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,2560,0.008309333688682979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,10240,0.024538666009902954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,7168,0.008960888617568547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,128,0.015446222490734525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,3072,0.02384977704948849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,2048,0.025230222278171118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,4096,0.017487999465730455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,7168,0.010384889112578498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,3584,0.007663999994595845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,2560,0.007277333074145847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,4096,0.011472889118724398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,1024,0.003169777699642711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,768,0.0030817778574095834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,32,0.009119110802809397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,12288,0.03253066539764404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,16384,0.020562666985723708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,10240,0.09703289137946235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,7168,0.035587555832333036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,12288,0.019911999503771465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,2560,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,64,0.011071999867757162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,12288,0.01754488878779941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,128,0.015206222732861837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,8192,0.01074044406414032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,1024,0.007632888853549957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,64,0.001830222292078866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,10240,0.01571999986966451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,3584,0.02626755502488878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,12288,0.04598755637804667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,2560,0.007280000381999546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,7168,0.04277155465549893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,65536,0.25626222292582196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,1536,0.010678222609890832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,7168,0.030299554268519085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,3072,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,1536,0.10364622539944118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,65536,0.36980001131693524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,10240,0.05267911156018575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,2560,0.006973333656787872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,7168,0.0321760012043847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,768,0.003034666594531801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,1536,0.016739555531077914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,128,0.020631111330456205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,4096,0.003951111187537511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,16384,0.02086399992307027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,64,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,12288,0.056010663509368896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,5120,0.04264000058174133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,8192,0.010795555180973478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,3584,0.009191110730171204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,256,0.017955554856194388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,16384,0.07835200097825792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,65536,0.06254488892025419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,1536,0.0035537779331207275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,16384,0.007992888490358988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,7168,0.013445333474212222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,5120,0.009342222577995723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,1536,0.004413333204057482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,512,0.0024764444679021835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,768,0.00793688909875022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,12288,0.01776088939772712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,256,0.011800888511869641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,1024,0.002924444360865487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,5120,0.017166222135225933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,2048,0.006586666736337874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,3072,0.016542222764756944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,1536,0.004150222159094281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,10240,0.04796977837880453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,32,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,2048,0.007270221908887227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,2048,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,1024,0.0034924443397257063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,4096,0.030812445614073012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,768,0.0029768887907266617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,512,0.005377777748637729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,64,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,1536,0.007992888490358988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,768,0.003208000005947219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,768,0.003945777813593547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,12288,0.05168177684148153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,2048,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,10240,0.022639999787012737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,64,0.0031635556370019913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,3072,0.005327111317051782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,1536,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,128,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,16384,0.020117332537968952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,512,0.014318222800890604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,256,0.03615199857287937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,1536,0.01579466627703773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,16384,0.046458668178982206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,1536,0.028775999943415325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,128,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,3072,0.009326222042242685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,128,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,256,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,1536,0.02014666630162133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,2560,0.009335111412737105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,65536,0.4021635585361057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,1536,0.022252443763944838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,256,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,128,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,64,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,3072,0.010437332921557956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,16384,0.02106399999724494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,1024,0.015504888362354703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,256,0.005199111170238919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,7168,0.00889866633547677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,8192,0.04009866714477539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,7168,0.018208889497650992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,4096,0.01963555481698778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,512,0.005647999958859549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,10240,0.034381333324644305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,3072,0.008644444247086843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,2560,0.011495999991893768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,10240,0.012065778175989786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,32,0.0018817778262827133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,2048,0.01349688900841607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,65536,0.0625457763671875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,6144,0.010718222293588849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,12288,0.0226248883538776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,3584,0.048453334305021495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,1024,0.017602667212486267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,12288,0.017055999901559617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,5120,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,3072,0.030908445517222088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,3072,0.010038221875826517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,5120,0.020288000504175823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,128,0.002179555594921112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,12288,0.06959466801749335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,512,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,1536,0.03992266787423028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,1024,0.007251555720965068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,64,0.007990222010347579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,10240,0.012084444363911947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,7168,0.021145777569876775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,10240,0.0323884454038408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,512,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,7168,0.019670221540662978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,6144,0.09526489178339641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,12288,0.036779556009504534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,12288,2.719416936238607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,16384,0.10536266697777642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,65536,0.2041768895255195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,2048,0.007239110767841339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,2048,0.09250844187206692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,768,0.0686195558971829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,7168,0.02648799949222141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,64,0.005980444451173146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,4096,0.009740444521109263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,3072,0.007183999651008182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,10240,0.018031110366185505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,64,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,5120,0.05837066968282064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,7168,0.20212977462344697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,1024,0.0028568889117903183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,256,0.015798222687509324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,6144,0.027786665492587622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,2560,0.030455999904208716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,8192,0.01608444419172075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,16384,3.8182996114095054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,65536,0.07194222344292535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,128,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,128,0.02164622147878011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,10240,0.014548444085650973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,256,0.008295111358165741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,3584,0.011744000017642975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,10240,0.042149331834581166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,16384,0.02271288964483473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,4096,0.31252890162997776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,1536,0.009674666656388177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,65536,0.14280445045895046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,256,0.006629333313968446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,7168,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,65536,0.3073964383867052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,1024,0.01014311114947001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,2560,0.017550221747822232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,16384,0.030921777089436848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,7168,0.014444443914625378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,16384,0.7871146731906467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,3584,0.02126844392882453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,128,0.008982222113344405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,16384,0.03382577829890781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,2048,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,6144,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,512,0.008274666965007782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,256,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,7168,0.010094222095277574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,512,0.008329778081840938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,8192,0.03184088733461168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,128,0.015111999379263984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,10240,0.058393776416778564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,1024,0.01484444406297472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,768,0.003846222327815162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,512,0.0025351111673646504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,2560,0.004934222333961063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,3584,0.03907377852333917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,3072,0.04773777723312378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,16384,1.5146702660454645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,10240,0.015145777000321282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,1536,0.009527111219035255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,6144,0.010703110860453712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,1024,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,1536,0.008492444124486711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,12288,0.05005777875582377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,1024,0.01889155473974016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,128,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,65536,0.18824711110856798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,768,0.017158221867349412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,2048,0.007296889192528195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,768,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,5120,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,768,0.01755822201569875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,128,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,32,0.0083155557513237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,10240,0.026477333572175767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,3584,0.007635555333561367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,1024,0.015493333339691162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,6144,0.011710222396585675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,256,0.016169778174824186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,512,0.02553066611289978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,1536,0.028464890188641016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,3584,0.2051173316107856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,256,0.016182222300105624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,8192,0.06346844302283393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,3584,0.00703022215101454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,5120,0.00867911097076204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,5120,0.01834222177664439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,10240,0.004872888740566042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,3072,0.01814399990770552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,3584,0.008857777549160851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,512,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,4096,0.010760000182522668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,2560,0.006952889263629913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,256,0.006635555376609166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,6144,0.016516443755891588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,7168,0.20510933134290907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,1024,0.00798044436507755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,3072,0.008284444610277811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,16384,0.060356444782680936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,10240,0.07032533486684163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,16384,0.08950666586558025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,3584,0.011049778097205691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,1536,0.009353777600659264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,256,0.01680888897842831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,1536,0.009002666506502364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,4096,0.0632622241973877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,1536,0.02887644370396932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,65536,0.06268977456622653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,16384,0.054140445258882314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,16384,0.04548177785343594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,128,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,6144,0.037920888927247785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,512,0.013406222065289816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,1536,0.025695110360781353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,2560,0.011694221860832639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,10240,0.014724445011880664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,512,0.0069573331210348345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,1536,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,65536,0.09462222125795151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,16384,0.02106133268939124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,256,0.002479111154874166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,10240,0.028359999259312946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,3072,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,8192,0.02574577762020959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,2048,0.008651555412345463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,768,0.01756000022093455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,4096,0.03388444582621256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,8192,0.022610666023360357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,1024,0.004519999855094486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,3584,0.01720266706413693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,12288,0.021584888299306233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,3584,0.00959822204377916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,12288,0.12042044268714057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,3072,0.005664888769388199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,768,0.007672000262472365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,3072,0.007633777956167857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,16384,0.023592889308929443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,5120,0.10788977808422512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,4096,0.01697422232892778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,10240,0.01714666684468587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,65536,0.13559288448757595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,512,0.020896888441509668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,16384,0.03897422220971849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,512,0.007624000310897827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,512,0.005376889059940974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,3584,0.018605333235528734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,10240,0.013264888690577613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,7168,0.005533333453867171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,10240,0.04799911048677233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,256,0.017163554827372234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,2048,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,3072,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,10240,0.027007111244731482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,64,0.0052622221410274506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,65536,0.06289600001441108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,256,0.015114666687117683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,7168,0.011750222080283694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,16384,0.03613422314325968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,10240,0.08857778045866226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,512,0.002471999989615546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,256,0.0069217777086628815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,10240,0.014765333798196582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,12288,0.08945244550704956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,3584,0.02295822236273024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,12288,0.025428444147109985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,128,0.01515111161602868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,2048,0.004517333375083076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,768,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,128,0.0018026667336622875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,2560,0.012359999948077731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,3584,0.014848000473446317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,8192,0.08180444770389132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,4096,0.008631111019187504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,12288,0.08480355474683975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,128,0.010335110955768162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,12288,0.03477511141035292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,12288,0.11434933874342178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,8192,0.015820443630218506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,2560,0.010415111151006486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,5120,0.0511813329325782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,5120,0.011335111326641507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,4096,0.024440889557202656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,2560,0.006976000136799282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,10240,512,0.01037333326207267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,768,0.007250666618347168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,4096,0.005961777849329843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,5120,0.021294222937689886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,12288,0.06274666388829549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,12288,0.014831999937693277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,768,0.013826666606797112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,10240,0.21441511313120523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,64,0.001806222316291597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,12288,0.03088800112406413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,4096,0.0074026667409473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,16384,0.036157333188586764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,512,0.01615111033121745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,1024,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,16384,0.136318220032586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,7168,0.009713778065310584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,5120,0.008045333127180735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,128,0.00250133333934678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,6144,0.01998222205373976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,5120,0.00867022242810991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,10240,0.06754755311542086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,7168,0.1329475508795844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,10240,0.015835555063353646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,2048,0.009296889106432596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,256,0.0021679999513758556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,6144,0.02000799940692054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,12288,0.03125955661137899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,2048,0.03812355465359158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,768,0.016442666451136272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,1024,0.0035217776894569397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,10240,0.01481244464715322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,768,0.034097777472601995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,1024,0.003843555433882607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,128,0.009352888498041365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,6144,0.009344889058007134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,768,0.02795555525355869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,6144,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,6144,0.00976533359951443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,10240,0.01758044461409251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,3584,0.015820443630218506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,128,0.021632888250880774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,3584,0.028443555037180584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,10240,0.033766223324669734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,1536,0.03344800074895223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,2560,0.01315822203954061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,3584,0.00831288927131229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,256,0.0022311111291249595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,2560,0.004942222187916438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,512,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,256,0.012416000167528788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,7168,0.011157333023018308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,2048,0.0069715554515520734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,3584,0.007992000215583378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,7168,0.044461333089404635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,12288,0.019668444991111755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,1024,0.005962666538026597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,5120,0.016511999898486666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,6144,0.013880000346236758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,512,0.023356444305843774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,768,0.013749333719412485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,4096,0.01758399936887953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,3072,0.021020443903075323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,16384,0.02382933265633053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,32,0.0021217777911159727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,3584,0.01717155509524875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,2048,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,4096,0.014667555689811707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,512,0.005579555614127054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,2048,0.014115555418862237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,2048,0.009291555318567488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,16384,0.07301866345935397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,1536,0.005934222290913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,65536,0.05948355462816027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,256,0.002465777720014254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,6144,0.02595377796226078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,2048,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,65536,0.2050257788764106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,768,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,65536,0.14491732915242514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,128,0.006330666856633292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,65536,0.10025066799587673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,12288,0.032487998406092324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,1024,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,2048,0.03794044587347243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,10240,0.02867377797762553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,3072,0.017231111725171406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,32,0.00591555568906996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,8192,0.034809778134028115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,8192,0.03849155704180399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,7168,0.006952889263629913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,128,0.014488889111412896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,768,0.013737777868906656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,3072,0.038775109582477145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,6144,0.010341333018408881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,2048,0.01404711107412974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,65536,0.05097600155406528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,16384,0.01497244503762987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,7168,0.016410667035314772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,64,0.007552000383536021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,12288,0.018389332625601027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,32,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,2048,0.0042257776690853965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,512,0.0028373334142896864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,5120,0.02440888848569658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,12288,0.2725253370073107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,2048,0.006781333436568578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,4096,0.02478844424088796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,5120,0.006636444479227066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,2560,0.012463111016485425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,8192,0.024528889192475214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,5120,0.016041777200169034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,65536,0.07547022236718072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,768,0.006500444478458828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,256,0.0090506664580769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,10240,0.05916355715857612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,2048,0.009022222624884712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,12288,0.013535999589496188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,4096,0.05722666449016995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,10240,0.030155556069480047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,5120,0.05079288946257698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,8192,64,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,10240,0.047599110338422984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,1024,0.019559999306996662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,1536,0.008289777570300633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,10240,0.018604444132910836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,6144,0.008667555948098501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,5120,0.028907554017172918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,6144,0.060144000583224826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,16384,0.039874666266971164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,256,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,1024,0.015132443772421943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,128,0.006960888703664144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,12288,0.03511733478969998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,3072,0.049710220760769315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,10240,0.015751111838552687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,1536,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,768,0.006078222145636876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,2048,0.008279999925030602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,7168,0.0537048876285553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,2048,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,6144,0.041639112763934664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,64,0.0019004444281260173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,256,0.006188444379303191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,32,0.0021137777302000257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,2048,0.0063866666621632045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,5120,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,4096,6144,0.011385777758227454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,1536,0.0035439998739295532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,256,0.02295644415749444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,65536,0.09366133477952744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,3584,0.010493333141009012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,4096,0.008287111090289222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,2048,0.01481599940194024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,10240,0.02419288953145345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,7168,0.00998044427898195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,2048,0.007491555478837755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,6144,0.11373421880933975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,3584,0.37776801321241593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,7168,0.039276444249682956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,8192,0.06251644425921969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,1536,0.0034684443639384378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,10240,0.024825778272416856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,10240,0.012080888781282636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,3584,0.008217777642938826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,512,0.014512888259357877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,5120,0.013445333474212222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,1536,0.06849777698516846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,512,0.006253333141406377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,1536,0.012792000340090858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,6144,0.012986666626400419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,512,0.005951111101441913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,512,0.0058782220714622075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,512,0.005495999836259418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,6144,0.02093688812520769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,2560,0.10218489170074463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,768,0.0028533333291610083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,7168,0.028239111105600994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,768,0.017887999614079792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,128,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,16384,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,8192,0.01072266697883606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,256,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,3584,0.007479999628331926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,7168,0.030864887767367895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,5120,0.009992000129487779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,5120,0.00923911150958803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,256,0.0032008888406885993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,12288,0.047637333472569786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,512,0.010002666877375709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,4096,0.08992089165581597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,10240,0.03940533267127143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,65536,0.15753155284457737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,32,0.0021173333128293357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,5120,0.20765688684251574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,65536,0.06804355647828844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,6144,0.05462755428420173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,65536,0.21463110711839464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,2048,0.03387377659479777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,8192,0.018868444694413077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,1536,0.0764933360947503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,5120,0.012704888979593912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,256,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,1536,0.09088266558117336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,2048,0.0038817777401871155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,4096,0.007248889240953658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,10240,0.00759733302725686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,7168,0.020486222373114694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,16384,0.17325333754221597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,10240,0.03848000036345588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,4096,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,4096,0.008967111508051554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,2560,0.01920622256067064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,10240,0.013391999734772576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,768,0.021191999316215515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,5120,0.10751022232903375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,3072,0.015610666738616096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,1536,0.008643555144468943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,4096,0.0340320004357232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,3072,0.0058746664888328975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,3584,0.08879199955198501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,6144,0.018238221605618794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,1536,0.003216888962520493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,3584,0.016137777103318107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,64,0.005123555660247803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,512,0.06647644440333049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,6144,0.01844800015290578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,64,0.023688889212078516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,3072,0.007323555648326874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,6144,0.031888888941870794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,4096,0.009386666946940953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,32,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,10240,0.03874489002757602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,1536,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,3072,0.007812444534566667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,3584,0.020621332857343886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,10240,0.012058667010731168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,8192,0.010610666539933948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,1024,0.008584000170230865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,32,0.0017866666118303935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,3072,0.07228000296486749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,3072,0.007623111208279927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,256,0.03424266642994351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,8192,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,7168,0.02014400064945221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,7168,0.030838221311569214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,2048,0.004206222378545337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,128,0.030596445004145306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,5120,0.04191199938456217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,4096,0.00796622203456031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,5120,0.017850667238235474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,128,0.014506666196717156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,5120,0.3486622174580892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,8192,0.027287999788920086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,16384,0.01275111072593265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,8192,0.02940000096956889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,64,0.00831288927131229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,4096,0.007931555310885111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,65536,0.06269155608283149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,7168,0.019842666056421068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,10240,0.015223110715548197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,3584,0.03463644451565213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,8192,0.010727110836240979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,65536,0.07768978012932672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,128,0.014893333117167154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,8192,0.015423110789722867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,7168,0.026763555076387193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,256,0.010825778047243753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,2560,0.02706577711635166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,8192,0.039684444665908813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,1024,0.0036151111125946045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,6144,0.014720888601409065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,12288,0.02735022207101186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,64,0.0021342221233579847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,256,0.020621332857343886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,1024,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,2560,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,1536,0.10174133380254109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,2560,0.018629334039158292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,2560,0.011555555793974133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,1024,0.006269333263238271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,64,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,8192,0.010990222295125326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,1024,0.00590844452381134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,768,0.006205333189831839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,3072,0.007161777880456712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,512,0.004435555388530095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,7168,0.0734586649470859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,4096,0.010481778118345471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,2560,0.013808888693650564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,5120,0.00453333349691497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,256,0.016568000117937725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,64,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,768,0.018735110759735107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,10240,0.060159113672044545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,256,0.016855110724767048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,8192,0.04310933417744107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,6144,0.1461084418826633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,3584,0.01163200040658315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,8192,0.0503057771258884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,64,0.01723911033736335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,3584,0.024290665984153748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,512,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,4096,0.008260444634490544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,12288,0.031159109539455835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,512,0.005392000079154968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,8192,0.5420328776041666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,32,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,3584,0.009686222506894005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,8192,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,10240,0.06073066923353407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,768,0.020584889584117465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,4096,0.012109333442317115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,256,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,128,0.0031742221779293488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,4096,0.018894222047593858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,4096,0.04523377617200216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,2560,0.03186755710177951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,128,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,65536,0.022652443912294176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,7168,0.013139555851618448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,16384,0.015816888875431485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,12288,0.013145777914259167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,3072,0.02162133322821723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,6144,0.01036888857682546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,256,0.014799111419253878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,768,0.0025031111306614345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,512,0.014753777119848462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,16384,0.01890222231547038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,1536,0.026769777139027912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,256,0.016861332787407767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,8192,0.01275111072593265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,7168,0.011700444751315646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,2048,0.006220444622966979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,3072,0.02718399961789449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,32,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,256,0.004200889004601372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,1536,0.00630044440428416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,6144,0.020003555549515616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,1536,0.005299555758635203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,1536,0.022854222191704646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,512,0.05618933174345228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,65536,0.06979733043246798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,12288,0.01531644496652815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,512,0.018887110882335238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1536,1024,0.0031884445084465873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,1536,0.010060444474220276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,16384,0.049104889233907066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,8192,0.021887110339270696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,65536,0.09760799672868516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,65536,0.22730933295355904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,3584,0.0337066650390625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,3072,0.013472889032628802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,4096,0.021661332911915247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,10240,0.02095199955834283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,32,0.011752000285519494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,1024,0.008012444608741337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,2048,0.0059164443777667145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,2048,0.006595555692911148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,1536,0.004587555511130227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,4096,0.06664533085293241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,10240,0.013745778136783175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,32,0.0052471111218134565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,768,0.004212444441186057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,3584,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,512,0.006190222170617845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,128,0.00646666685740153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,2560,0.028074665202034846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,3584,0.0052675555149714155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,512,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,16384,0.008915555146005418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,65536,0.11591377523210313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,128,0.00796088907453749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,10240,16384,0.10031555758582221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,128,0.026611554953787062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,1536,0.015207111835479736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,5120,0.021517333057191636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,16384,0.018251554833518136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,1024,0.05956355730692545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,16384,0.04684977730115255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,65536,0.13583377997080484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,12288,0.01684888866212633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,1536,0.010675555302037133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,3072,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,1536,0.02230577833122677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,128,0.015779554843902588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,5120,0.02981333269013299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,3072,0.022396443618668452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,12288,0.07464177740944757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,128,0.0020862221717834473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,4096,0.011743110915025076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,1536,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,1536,0.011956444217099084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,2048,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,5120,0.006566222343179915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,2048,0.00792711145348019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,7168,0.02384000023206075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,3072,0.011002666420406766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,5120,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,8192,0.005205333232879639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,6144,0.005646222167544895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,64,0.0041724443435668945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,10240,0.0474639998541938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,6144,0.022553778356975977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,65536,0.10268711381488377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,2560,0.004859555512666702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,12288,0.027109333210521277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,1024,0.06142222219043308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,3072,0.009372444616423713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,3072,0.03596444593535529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,768,0.006603555546866522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,6144,0.03690222236845229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,6144,0.018009778526094224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,5120,0.011788444386588203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,7168,0.011309333145618439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,768,0.003548444559176763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,3072,0.012929778132173749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,6144,0.01861422260602315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,12288,0.046163555648591786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,5120,0.025224000215530396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,65536,0.2660568820105659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,32,0.002144888871245914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,3072,0.007582222421964009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,512,0.0028133332315418455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,512,0.010006222460005019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,6144,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,1536,0.007127111156781514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,8192,0.019909333851602342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,2048,0.012096000214417776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,1024,0.156640887260437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,1024,0.00870311094654931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,12288,0.036814222733179726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,1536,0.004587555511130227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,7168,0.20474310715993246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,128,0.0022026666750510535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,768,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,6144,0.01930488811598884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,4096,0.01149511088927587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,3072,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,6144,0.023436443673239812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,2560,0.0070497774415545994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,6144,0.006976000136799282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,1536,0.015798222687509324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,256,0.005863111052248213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,512,0.0024862223201327855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,1536,0.004535111288229625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,1024,0.0041786668201287585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,2048,0.007364444434642792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,2048,0.00543200017677413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,512,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,3072,0.0888320008913676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,3584,0.019302222463819716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,7168,0.017228444417317707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,3584,0.009655999640623728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,256,0.01685599982738495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,512,0.013396444420019785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,1024,0.006322666588756774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,128,0.006670222100284364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,12288,6.962423960367839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,128,0.0567164421081543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,8192,0.016640888320075143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,5120,0.016347555650605094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,128,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,1024,0.005950222412745158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,2560,0.010778666370444827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,2048,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,256,0.006671111202902264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,128,0.014706666270891825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,128,0.014466666513019137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,768,0.0032231110251612137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,16384,0.021006221572558086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,6144,0.024347555306222703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,256,0.014785778191354541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,4096,0.04039644532733493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,4096,0.2006533278359307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,16384,0.2361848884158664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,16384,0.018915555543369718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,128,0.009663999908500248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,2560,0.010606221854686737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,3072,0.016884444488419425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,12288,0.037993778785069786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,3072,0.007363555332024892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,8192,0.013794666363133324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,512,0.007300444775157505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,6144,0.02590755621592204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,1024,0.02847466535038418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,3072,0.008172444171375698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,4096,0.008630221916569604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,128,0.015501333607567681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,128,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,16384,0.2482302188873291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,4096,0.007996444072988298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,3072,0.011404444773991903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,128,256,0.011013333168294696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,10240,1.8426221211751301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,12288,0.032073779238594904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,256,0.0162773331006368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,8192,0.021342222889264423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,2560,0.02164711058139801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,32,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,5120,0.10467644532521565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,128,0.002143111079931259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,5120,0.017660443981488545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,64,0.0021280000607172647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,12288,0.05319377779960632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,7168,0.05201955636342367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,64,0.002471999989615546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,2560,0.012633778154850006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,128,0.015092444088723926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,10240,0.023009777069091797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,65536,0.03350044290224711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,10240,0.02648177742958069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,7168,0.030172444052166406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,16384,0.08452888992097642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,128,0.0058764442801475525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,512,0.02201955517133077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,3072,0.00721599989467197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,512,0.0026693333768182327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,768,0.00282577777074443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,1536,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,3072,0.00797333319981893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,5120,0.006909333169460297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,512,0.056884447733561196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,4096,0.023129777775870428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,256,0.01274222218328052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,10240,0.026445334156354267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,128,0.011667555405033959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,10240,0.01831466621822781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,1536,0.01819022165404426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,1024,0.0603733327653673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,2560,0.08251200119654338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,65536,0.09004533290863037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,6144,0.020273778173658583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,5120,0.020625778370433383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,256,0.002136888810329967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,512,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,7168,0.013243555194801755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,4096,0.009327111144860586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,65536,0.07589333587222628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,32,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,256,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,10240,0.014241778188281588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,1024,0.030653334326214258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,3584,0.20115555657280815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,7168,0.13016266292995876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,1024,0.014936889211336771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,10240,0.0188746667570538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,7168,0.020615110794703167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,512,0.0032195556494924757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,128,0.015102222561836243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,32,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,2048,0.008933333059151968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,3072,0.021436444587177698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,1024,0.007714666426181793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,3584,0.03422755665249295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,65536,0.07247466511196561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,3584,0.008075555165608725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,768,0.0075777777367168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,2560,0.02500711050298479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,5120,0.031930665175120033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,768,0.03391022152370877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,2048,0.013450666434235044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,3072,0.004582222137186262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,3584,0.008638222184446123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,512,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,8192,0.0508266654279497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,64,0.0041617775956789655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,512,0.022631999519136217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,768,0.0027928888383838865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,10240,0.020272889071040683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,768,0.006246222390068903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,768,0.01683022247420417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,5120,0.02311022248533037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,256,0.016144000821643405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,10240,0.025705777936511572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,4096,0.008394666843944127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,1536,0.006077333456940121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,32,0.005873777800136142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,3584,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,5120,0.11617244614495172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,6144,0.04366133279270596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,768,0.007263110743628607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,768,0.01682133310370975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,1536,0.00638488887084855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,2560,0.01606666710641649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,3584,0.026439110438028973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,12288,0.05250755614704556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,4096,0.008253333469231924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,6144,0.0151591118839052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,65536,0.1427439981036716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,16384,0.047594666481018066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,1536,0.00758133331934611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,3072,0.09191911088095771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,65536,0.07942222224341498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,12288,0.02442577812406752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,10240,0.012777778009573618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,65536,0.5024684270222982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,512,0.005306666509972678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,256,0.02623822291692098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,3584,0.016547555724779766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,512,0.00351555562681622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,65536,0.33830756611294216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,4096,0.00868711123863856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,16384,0.07706399758656819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,1024,0.022620444496472675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,768,0.0028026666906144884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,1024,0.016530666086408827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,3072,0.03335377905103896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,7168,0.03126577867401971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,65536,0.26008444362216526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,1024,0.0035093331502543557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,32,0.002576888849337896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,64,0.004231111043029361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,16384,0.027094221777386133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,128,0.010397333237859938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,1024,0.0870293312602573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,10240,0.024138665861553613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,7168,0.021407110823525324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,768,0.0031964443624019623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,3584,0.010319111247857412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,16384,0.13751733303070068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,4096,0.011935110721323224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,2048,0.01613333324591319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,12288,0.019727110862731934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,64,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,256,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,2560,0.007438222567240397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,8192,0.5475520028008355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,6144,1.287740495469835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,6144,0.01826844447188907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,12288,0.01688888834582435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,5120,0.2744355466630724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,3072,0.007600000335110559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,10240,0.09150044123331706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,4096,0.03071911136309306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,12288,0.05739555756251017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,10240,0.022706665926509436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,32,0.005255110975768831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,65536,0.2222426732381185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,4096,0.007972444097201029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,10240,0.03570577833387587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,4096,0.019333332777023315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,768,0.015184000134468079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,6144,0.009337777892748514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,768,0.013463999662134381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,65536,0.6683982213338217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,5120,0.019144000278578866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,6144,0.05231022172504001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,6144,0.009577777650621202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,1024,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,1024,0.00319377767542998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,4096,0.026741334133678015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2048,512,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,3072,0.01585422290696038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,3072,0.018978666928079393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,512,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,256,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,16384,0.02492444382773505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,2048,0.02831022275818719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,3072,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,768,0.02290133304066128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,4096,0.02537600033813053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,3584,0.02054755555258857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,128,0.005071111023426056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,1536,0.00371111101574368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,5120,0.02552266584502326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,768,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,1536,0.006266666783226862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,65536,0.5854720009697808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,65536,0.47621244854397243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,2048,0.009296000003814697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,3584,0.026019554999139573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,768,0.003128000017669466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,4096,0.01571911076704661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,768,0.01721600029203627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,8192,0.6978631019592285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,7168,0.056242664655049644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,768,0.024102222588327196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,768,0.013773333695199756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,5120,0.010668444136778513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,8192,0.02591200007332696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,1024,0.009305777649084726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,7168,0.033336000310050115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,3584,0.010909332997269101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,7168,0.016938666502634685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,2048,0.028248889578713313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,8192,0.02295733326011234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,3072,0.010320000350475311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,32,0.0021422221842739317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,10240,0.016003555721706815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,2560,0.004908444566859139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,3072,0.02568622264597151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,256,0.0063822223908371395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,1536,0.008611555728647444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,10240,0.016851555969980028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,7168,0.11544177267286514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,65536,2.256024890475803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,8192,0.03710755705833435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,16384,0.8215164608425565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,768,0.0288702216413286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,32,0.010028444230556488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,2048,0.027335999740494624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,4096,0.033944000800450645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,3584,0.009346666435400644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,1024,0.014488889111412896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,8192,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,4096,0.2649279965294732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,32,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,32,0.0018026667336622875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,2048,0.016228444046444364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,7168,0.03805955582194858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3584,12288,0.5068711174858941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,6144,0.009663110805882348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,4096,0.008958222137557136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,5120,0.08795377943250869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,32,0.03626666797531976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,64,0.007359999749395583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,2560,0.016677333248986136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,12288,0.015478221906556023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,64,0.005894222193294101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,3072,0.00962400022480223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,128,0.020580444071027968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,16384,0.01641155613793267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,8192,0.007674666742483775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,3072,0.027496889233589172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,7168,0.1013759970664978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,1536,0.003822222352027893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,16384,0.14158311155107287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,768,0.003969777789380816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,16384,0.1284773349761963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,16384,0.0533884432580736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,8192,0.036421332094404436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,10240,0.0181057784292433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,65536,0.2410648928748237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,5120,0.026725333597924974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,1536,0.019259555472267997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,128,0.019905777441130746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,65536,0.21154756016201445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,3584,0.042167110575569995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,7168,0.016864889197879367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,8192,0.01239377756913503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,1024,0.003661333272854487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,8192,0.020954666866196528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,1536,0.01442488862408532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,10240,0.009561777942710454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,10240,0.012718222207493253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,32,0.0058746664888328975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,6144,0.028946667909622192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,16384,0.019101333287027147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,3584,0.024878222081396315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,8192,0.03484088844723172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,16384,0.015791111522250705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,16384,0.04760977625846863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,1024,0.015097777048746744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,256,0.005197333378924264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,128,0.009744000103738572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,6144,0.005205333232879639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,3072,0.025237333443429735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,10240,0.013070222404268054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,1536,0.015464888678656684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,1536,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,65536,0.2945306566026476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,7168,0.011295999917719098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,16384,0.021116443806224402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,1024,0.008735111190213097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,1536,0.009720000127951304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,768,0.017647110753589206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,2560,0.010394666757848527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,32,0.001814222170246972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,2560,0.00811644477976693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,5120,0.017902221944597032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,768,0.016484444340070088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,2048,0.013083555632167392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,768,0.008293333152929941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,3584,0.016125332978036668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,256,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,3584,0.007687110867765214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,3072,0.0052284445199701525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,3584,0.23235199186537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,3584,0.004875555634498596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,8192,0.014428445034556918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,3072,0.016864889197879367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,128,0.024841777152485315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,256,0.00815644446346495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,12288,0.01351199961370892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,4096,0.01645955608950721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,65536,4.725023905436198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,1536,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,3072,0.013721778160995908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,512,0.013798221945762634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,3584,0.01869422197341919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,5120,0.017047999633683097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,256,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,256,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,10240,0.028888887829250757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,1024,0.009694221946928237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,12288,0.04151555564668443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,16384,0.035028444396124944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,10240,0.036766221125920616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,6144,0.010006222460005019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,7168,0.04731555448638069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,256,0.007956444389290279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,1536,0.006084444622198741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,1536,0.014967999524540372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,16384,0.04371111260520088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,4096,0.02402399977048238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,10240,0.012924444344308643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,3584,0.01583377851380242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,4096,0.01850133306450314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,16384,0.03475822342766656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,65536,0.17249600092569986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,512,0.016186666157510545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,3072,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,512,0.0032231110251612137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,16384,0.04671022295951843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,2560,0.01036177741156684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,10240,0.012042666474978128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,2048,0.04260711206330193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,16384,0.061582220925225147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,7168,0.045831110742357045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,1024,0.006671999891599019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,2048,0.009015111459626092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,5120,0.016392888294325936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,4096,0.019119110372331407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,7168,0.025425778494940862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,8192,0.010455110834704505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,12288,0.014091556270917257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,768,0.0417431096235911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,128,0.007998222278224098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,512,0.012587555580668978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,4096,0.03287555442916022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,512,0.005164444446563721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,512,0.011718221836619906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,2560,0.018477333916558158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,1536,0.013103110922707452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,10240,0.045831110742357045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,32,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,768,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,5120,0.040896889236238264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,4096,0.009695111049546136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,4096,0.004593777573770947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,7168,0.013373333546850415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,2560,0.021329777108298406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,8192,0.011001778145631155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,2560,0.024454222785101995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,12288,0.04557777775658501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,4096,0.028171555863486394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,10240,0.21104089419047037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,1536,0.009967111051082611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,512,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,512,0.01854933301607768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,5120,0.00848711116446389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,4096,0.04343377881579929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,256,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,10240,0.22114045090145537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,2048,0.009361777868535783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,128,0.005230222311284807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,1536,0.011732444167137146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,4096,0.02551199992497762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,1536,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,5120,0.008362666600280339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,16384,0.31059198909335667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,2048,0.0042133331298828125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,12288,0.018179555733998615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,8192,0.08697955475913154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,12288,0.0996311108271281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,3072,0.02812888887193468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,10240,0.03531466590033637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,12288,0.013167111410035027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,7168,0.053921778996785484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,1536,0.008295999632941352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,2560,0.016777777009540133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,10240,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,3072,0.10239377948972915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,4096,0.008299555215570662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,3584,0.09087466531329685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,2560,0.011559999651379056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,1024,0.006321777900060018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,768,0.005909333212508096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,7168,0.027422222826215956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,3584,0.010924444430404238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,1536,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,65536,0.21463021967146131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,128,0.0147386673423979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,1536,0.0192239996459749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,65536,0.08775733576880561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,768,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,3072,0.00905777762333552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,16384,3.7730187310112844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,65536,0.06497244702445136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,3072,0.011948444777064852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,512,0.006149333384301927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,4096,0.0417537788550059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,10240,0.04481600059403313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,7168,0.14727289146847194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,65536,3584,0.03565244542227851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,10240,0.012099555797047086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,256,0.015522667103343539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,2048,0.004853333450025982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,512,0.005950222412745158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,8192,0.012390221986505719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,512,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,256,0.0052737775776121355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,4096,0.01605244477589925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,768,0.0028311111446883944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,768,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,3584,0.024076445235146418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,4096,0.01830844415558709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,4096,0.018073777357737224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,12288,0.017736888594097562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,512,0.002489777695801523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,1536,0.010370666782061258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,12288,0.05070844292640686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,1024,0.009864889085292816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,5120,0.009012444151772393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,4096,0.004563555535342959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,6144,0.009485333330101436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,12288,0.019331556227472093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,8192,0.028248889578713313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,2048,0.008956444760163626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,3072,0.02565866708755493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,7168,0.010320888625250923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,3072,0.012200888660218982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,128,0.005720888988839255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,8192,0.016656888855828177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,3584,0.007616000043021307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,6144,0.009935999910036722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,1536,0.034240000777774386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,16384,0.022011554903454248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,5120,0.05933600001864963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,768,0.0026364444444576898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,1536,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,1024,0.015261333849694995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,10240,0.44446224636501735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,3072,0.025742222865422566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,1024,0.003536888708670934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,512,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,65536,0.30947555436028373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,2560,0.007000889215204451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,2048,0.010732444624106089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,128,0.01479466590616438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,3584,0.029504001140594482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,12288,0.02474755545457204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,6144,0.044477333625157676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,5120,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,5120,0.009001777403884465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,6144,0.03323200013902452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,65536,0.05845866600672404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,6144,0.03735911183887058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,512,0.005231110999981563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,64,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,1024,0.0058746664888328975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,2048,0.03828800055715773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,8192,0.0298977792263031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,12288,0.10499911175833808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,768,0.01633155511485206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,6144,0.022639110684394836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,512,0.014521777629852295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,256,0.006276444428496891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,10240,0.06495110856162177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,1024,0.007654222349325816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,16384,0.043613334496816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,10240,0.01204177737236023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,768,0.0028106667515304354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,512,0.0028320000403457214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,5120,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,12288,0.03881511092185974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,256,0.002193777718477779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,1024,0.013773333695199756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,5120,0.017543110582563613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,4096,0.0141004439857271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,16384,0.03291733397377862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,768,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,4096,0.015138667490747241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,4096,0.024938666158252295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,512,0.009328000247478485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,3072,0.01516355574131012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,5120,0.1996071073744032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,1024,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,3584,0.023707555400000677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,16384,0.01943199998802609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,3072,0.009688888986905416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,2048,0.009166222479608323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,2048,0.007630222373538547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,4096,0.028020444843504164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,64,0.04763911167780558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,1536,0.0047128887640105355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,7168,0.009697777529557547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,768,0.005610666755172942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,768,0.005944889038801193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,3584,0.05606666538450453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,1024,0.008242666721343994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,1536,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,128,0.008005333443482717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,2560,0.009668444593747457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,2560,0.008995555341243744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,8192,0.04484622346030342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,7168,0.04617244336340162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,16384,0.04186044467820061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,1024,0.01444533301724328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,3072,0.007691555553012424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,7168,0.03873422079616123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,16384,0.01816444430086348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,6144,0.032235556178622775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,12288,0.017733333839310538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,1536,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,5120,0.01680799987581041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,128,0.005967999911970562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,2560,0.02743822170628442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,12288,0.012862222062216865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,10240,0.04930399854977926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,10240,0.024162666665183172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,1024,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,2560,0.049717333581712514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,4096,0.021301334102948506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,65536,2.5115163591172958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,10240,0.04835377799140083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,5120,0.012662221988042196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,6144,0.012758221891191272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,3584,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,5120,0.03351733419630263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,65536,0.06546577480104235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,6144,0.2611084514194065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,65536,0.05872622463438246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,1024,0.006432000133726332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,5120,0.03119822343190511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,65536,0.15410399436950684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,6144,0.021192888418833416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,16384,0.043656888935301036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,5120,0.05524444580078125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,16384,0.023317333724763658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,6144,0.01882488860024346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,10240,0.027624888552559748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,32,0.0018320000833935207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,6144,0.04095377855830722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,6144,0.013374221821626028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,256,0.0031395554542541504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,256,0.03839911023775736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,3584,0.008275555239783393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,64,0.005884444548024072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,5120,0.010431999961535135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,4096,0.010428444378905825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,2048,0.00962933318482505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,5120,0.04708710975117154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,5120,0.008646222452322641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,2560,0.01881866653760274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,6144,0.014765333798196582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,7168,0.01924888955222236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,64,0.005533333453867171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,3584,0.1319013304180569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,8192,0.32685332828097874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,4096,0.020188444190555148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,65536,0.19524978266821968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,1024,0.00868622213602066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,512,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,5120,0.0069146665434042616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,16384,0.05325422353214688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,1024,0.008357333640257517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,7168,0.010671999719407825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,5120,0.008647110727098253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,16384,0.01995822290579478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,2560,0.007303111255168915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,3584,0.007768889268239339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,16384,0.01947377787695991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,768,0.10769689083099365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,3584,0.007320000065697565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,8192,0.01570044457912445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,6144,0.004917333523432414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,256,0.0021315556433465746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,768,0.005612444546487596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,10240,0.05334666702482435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,10240,0.09184355868233575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,16384,0.030817776918411255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,7168,0.020810667011472914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,256,0.0069066666894488865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,5120,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,6144,0.037069333924187556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,12288,0.09002044465806748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,128,0.006555555595291986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,4096,0.0041911109454101985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,768,0.018221333622932434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,5120,0.014649778604507446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,7168,0.025798221429189045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,8192,0.0158942225906584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,768,0.0034942221310403612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,64,0.006980444822046492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,3072,0.013696889082590738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,128,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,12288,0.13654311498006186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,1024,0.019392000304328073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,512,0.006883555402358373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,1536,0.007288888924651676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,128,0.002471999989615546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,1024,0.004181333300140169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,256,0.01720266706413693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,65536,0.0630577802658081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,128,0.011033777561452655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,8192,0.0255377772781584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,5120,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,8192,0.011412444214026133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,512,0.006974221931563483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,4096,0.01129244433508979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,2560,0.15598755412631565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,768,0.01757244434621599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,32,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,3584,0.029923554923799302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,1024,0.0031884445084465873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,32,0.00554933316177792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,128,0.0032248888164758682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,10240,0.060398221015930176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,12288,0.035070220629374184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,12288,0.20066932837168375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,128,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,1536,0.00997155573632982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,8192,0.007084444165229797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,32,0.0025555555605226094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,5120,0.01479644411140018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,7168,0.009714666340086196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,8192,0.02126400007141961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,6144,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,7168,0.027315555347336665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,6144,0.023060444328520033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,2560,0.0103502223889033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,64,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,2048,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,1536,0.02442577812406752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,1536,0.022867555419603985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,768,0.0057848890622456866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,3584,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,64,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,8192,0.03969688879119025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,768,0.017157332764731515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,3584,0.004868444469239977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,32,0.00656355544924736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,1024,0.006261333409282897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,5120,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,6144,0.016103999482260812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,4096,0.007924444145626491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,6144,0.02184711065557268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,1024,0.0032195556494924757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,12288,0.014128888646761576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,3072,0.016678222351604037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,12288,0.041738665766186185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,3584,0.004911111046870549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,3072,0.03801244497299194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,7168,0.017151999804708693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,768,0.017888888716697693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,65536,0.14920088979932997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,2048,0.026032888227038916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,128,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,32,0.005647111270162795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,7168,0.024005333582560223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,768,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,2560,0.07935021983252631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,4096,0.03977600071165297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,256,0.002526222210791376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,8192,0.018191110756662156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,4096,0.02355555527740055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,2048,0.004924444274769889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,1536,0.01975288821591271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,32,0.007632888853549957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,5120,0.01242844429281023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,5120,0.008657777474986183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,12288,0.01943555474281311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,6144,0.009357333183288574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,6144,0.0177511109246148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,6144,0.007970666719807519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,1024,0.008017777568764156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,6144,0.026864889595243666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,2048,0.006364444477690591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,16384,0.01889155473974016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,5120,0.022643556197484333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,2048,0.013287111288971372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,512,0.002493333278430833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,16384,0.027173333697848853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,2560,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,10240,0.028755555550257366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,1024,0.006968888971540663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,7168,0.010667555862002902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,65536,4.786983913845487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,3072,0.025164445241292317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,16384,0.03117155697610643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,4096,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,10240,0.010372444159454769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,65536,0.011696000066068439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,128,0.016150222884284127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,32,0.0028657778683635923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,12288,0.18915643956926134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,1536,0.02402044501569536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,512,0.002451555596457587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,16384,0.02700622214211358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,2048,0.011800000237094032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,6144,0.13488622506459555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,6144,0.009177777502271865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,65536,0.07980177799860637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,1536,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,6144,0.01092266705301073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,512,0.00926577796538671
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,8192,0.01069777790043089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,3584,0.02200177808602651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,65536,0.06427022483613756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,3072,0.02029511166943444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,12288,0.021233777205149334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,6144,0.012695111334323883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,5120,0.011728888584507836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,10240,0.019926221834288705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,2048,0.02068444424205356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,512,0.012119111087587146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,128,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,2560,0.010318222145239512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,7168,0.012930666406949362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,10240,0.052263110876083374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,128,0.004943999979231092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,10240,0.015407111909654407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,4096,0.03583022289805942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,1536,0.00350222239891688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,4096,0.23429155349731445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,6144,0.02568088968594869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,256,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,128,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,768,0.0041982221106688184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,3072,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,512,0.0069448889957533936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,3584,0.016384000579516094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,512,0.0025555555605226094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,3072,0.00796888851457172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,1536,0.021054221524132624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,3584,0.00601333338353369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,512,0.012098666694429187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,10240,0.04380444354481167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,3072,0.03029600116941664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,3072,0.01717688971095615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,512,0.01236088905069563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,4096,0.08624621894624497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,2560,0.09510311153199937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,128,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,4096,0.008656000097592672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,2048,0.014503111441930136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,128,0.014510222607188754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,512,0.016682666209008958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,8192,0.01166044423977534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,3072,0.013004444539546967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,3072,0.009175111022260454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,7168,0.08047378063201904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,10240,0.09414133098390366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,256,0.0021555556191338436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,3072,0.08449066347546047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,5120,0.42287643750508624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,2560,0.02067733307679494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,8192,1.0666942596435547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,2560,0.011344888971911537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,1024,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,2560,0.010936888555685679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,10240,0.016522667474216886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,1536,0.0036728887094391715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,12288,0.06562844249937269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,12288,0.04389422138532003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,1024,0.02606222199069129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,10240,8.29268561469184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,65536,0.15237332714928523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,256,0.009768889182143742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,1024,0.005877333382765452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,6144,0.059283554553985596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,512,0.05659288830227322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,32,0.0032026666320032547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,10240,0.016968000266287062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,512,0.01309866706530253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,3584,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,3584,0.014152889450391134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,4096,0.016954667038387723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,12288,0.7081182267930773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,12288,0.0558168888092041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,10240,0.020957332518365648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,65536,0.07981244723002116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,1536,0.01646577815214793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,256,0.007321777443091075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,32,0.008674666285514832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,5120,0.08643555641174316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,10240,0.012368000215954252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3584,16384,0.036327110396491155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,5120,0.00865955568022198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,16384,0.021756443712446425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,1536,0.003567999849716822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,2560,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,4096,0.010767111347781287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,32,0.001826666709449556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,6144,0.026615111364258662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,3584,0.026774222652117412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,3584,0.02986933456526862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,512,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,128,0.0028275555620590844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,1536,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,512,0.003522666792074839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,10240,0.017089777522616915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,5120,0.014109333356221518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,8192,0.011009777585665384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,16384,0.0546320014529758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,6144,0.010253333383136326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,2560,0.029114666912290785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,128,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,2560,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,512,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,1024,0.038077334562937416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,4096,0.011335999601417117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,2048,0.01942933268017239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,3584,0.008007110820876228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,768,0.01994400057527754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,4096,0.010496888723638324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,10240,0.028004444307751123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,1536,0.014751111467679342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,32,0.06845866971545748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,4096,0.02596355477968852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,6144,0.07125244537989299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,8192,0.014431110686726041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,16384,0.020255110330051847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,64,0.06772710879643758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,768,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,4096,0.009316444396972656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,3072,0.006795555353164673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,10240,0.03558933403756883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,3584,0.00685777763525645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,2048,0.009881777895821465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,4096,0.014357333381970724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,768,0.018244443668259513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,12288,0.039735111925337054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,16384,0.03159288896454705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,5120,0.03525866733656989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,7168,0.01942399972014957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,2048,0.011531555818186866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,3584,0.008000888758235509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,5120,0.016155555844306946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,5120,0.004217777815130022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,6144,0.039451556073294744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,6144,0.010377777947319878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,1024,0.005653333332803514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,65536,0.059360000822279185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,4096,0.013349333571063148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,32,0.0018151110659042995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,2560,0.012366222010718452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,6144,0.0369795560836792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,2048,0.010864888628323873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,8192,0.022286223040686712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,256,0.005192888693677055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,7168,0.2921911080678304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,3584,0.03128533230887519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,3072,0.007663999994595845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,65536,0.08299377891752455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,768,0.015104000767072042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,7168,0.03491910960939195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,6144,0.033391998873816595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,512,0.016511999898486666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,768,0.01649599936273363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,512,0.009283555878533257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,256,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,768,0.0057653333577844836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,512,0.01720266706413693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,32,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,4096,0.008168000313970778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,1536,0.00396799999806616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,6144,0.04242488741874695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,8192,0.016183111402723525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,4096,0.024143111374643113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,12288,0.01791288952032725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,2560,0.007277333074145847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,8192,0.015199111567603217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,10240,0.07869778076807658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,256,0.004983110974232356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,16384,3.147794723510742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,6144,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,1024,0.0041680000722408295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,65536,0.381881766849094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,1024,0.007958222594526079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,256,0.018215111560291715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1024,64,0.0070675553547011475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,8192,0.022608889473809138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,64,0.006953777538405524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,16384,0.027535999814669292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,12288,0.07011644707785712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,32,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,5120,0.11638043986426459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,2048,0.007963555554548899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,6144,0.09001066949632432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,128,0.01145866678820716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,512,0.04609422220124138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,3072,0.07166222068998548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,8192,0.053156445423762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,128,0.00589155571328269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,64,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,10240,0.02458400030930837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,8192,0.01478844384352366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,64,0.0024871110088295406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,4096,0.04526399903827243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,8192,0.06771110826068454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,10240,0.045144889089796275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,256,0.008488888541857401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,1536,0.006266666783226862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,5120,0.05110133356518216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,1024,0.00868977771864997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,5120,0.012162666353914471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,128,0.00526311124364535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,768,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,6144,0.008633777499198914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,4096,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,12288,0.023362666368484497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,1024,0.012083555261294046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,3072,0.007344000041484833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,3072,0.010280888941552904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,3072,0.024660444921917383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,4096,0.007481777833567725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,6144,10240,0.03366044494840834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,128,0.004179555508825514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,8192,0.01239733315176434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,6144,0.009646221995353699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,3584,0.009361777868535783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,5120,0.013189333180586496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,768,0.006160000132189856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,10240,0.04833244283994039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,12288,0.01610933409796821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,3584,0.016511110795868766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,12288,0.011758222348160215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,8192,0.010981332924630908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,768,0.005625777774386936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,1536,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,8192,0.04543911086188423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,3072,0.06227644284566244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,1536,0.022545778089099463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,12288,0.03543466660711501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,1536,0.00424888895617591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,7168,0.03527644607755873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,512,0.008303999900817871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,512,0.002502222235004107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,768,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,512,0.014597333139843412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,2048,0.005666666560702854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,65536,0.08095910814073351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,8192,0.3857431146833632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,65536,0.06919911172654894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,128,0.012439111040698158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,8192,0.028416888581381902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,3072,0.00830844458606508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,64,0.0021502222451898786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,1536,0.011023999916182624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,1024,0.006794666664467917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,4096,0.024471999870406255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,1536,0.0038942222793896994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,65536,0.07487555344899495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,10240,0.026018665896521673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,8192,0.1157875590854221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,65536,0.2301102214389377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,16384,0.05498577819930183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,2560,0.03455555438995361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,10240,0.009673777553770278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,2048,0.004879111217127906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,7168,0.011019555230935415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,3584,0.061556445227728955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,16384,0.010828444527255164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,6144,0.02188266648186578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,6144,0.03091644578509861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,256,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,5120,0.011810666984981961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,512,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,3072,0.035707556539111666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,1536,0.004690666579537922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,256,0.014900444282425774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,65536,0.13699466652340359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,4096,0.01036622209681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,3072,0.03930755456288656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,2048,0.017521777086787753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,2048,0.0069653333889113525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,1536,0.01533511115445031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,16384,0.016474665866957772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,512,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,512,0.014094221923086377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,2048,0.02830666634771559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,7168,0.03160888950030009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,256,0.009350222018029954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,3072,0.6619111167060004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,8192,0.13160622119903564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,2048,0.02125066684352027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,4096,0.00452622233165635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,64,0.01035288886891471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,3072,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,1024,0.0047351109484831495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,128,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,5120,0.027432888746261597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,4096,0.008281778130266402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,768,0.005574222240183089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,7168,0.02029244436158074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,2560,0.01648355523745219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,6144,0.014537778165605334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,2560,0.02661955522166358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,5120,0.019834667444229126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,1536,0.006636444479227066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,128,0.012076444096035428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,1536,0.01864622202184465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,10240,0.01311555587583118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,768,0.020454222957293194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,1536,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,16384,0.020011555817392137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,1536,0.0185199992524253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,6144,0.011619555453459421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,7168,0.039697776238123574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,12288,0.016912889149453905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,2560,0.006686222222116258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,16384,0.028415110376146104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,128,0.015840889679061044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,128,0.005952888892756567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,65536,0.24359732204013398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,16384,0.03530044356981913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,2048,0.014427555931939019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,5120,0.028756442997190688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,4096,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,1536,0.0069857777820693115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,5120,0.03152088986502753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,12288,0.105731553501553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,768,0.005955555372767978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,7168,0.02072355482313368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,16384,0.05084977878464592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,4096,0.00868622213602066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,8192,0.01817599932352702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,16384,0.005937777873542573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,10240,0.016353777713245816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,128,0.01477866702609592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,65536,0.12255644798278809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,4096,0.007646222081449296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,2560,0.023676445086797077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,5120,0.010066666536860997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,768,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,64,0.006852444261312485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,128,0.014470222923490735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,3072,0.007667555577225155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,1024,0.00398577791121271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,1536,0.011525332927703857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,2048,0.04653066727850172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,65536,0.16505332787831625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,16384,0.017140444782045152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,1024,0.002817777709828483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,3584,0.017588444881969027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,1536,0.02216888964176178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,512,0.013414222333166333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,65536,0.0647066699133979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,4096,0.015124445160230001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,7168,0.08044622341791789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,512,0.005063999858167437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,3072,0.011324444578753578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,512,0.0024666666156715816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,8192,0.024298666252030268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,32,0.004520888957712385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,768,0.023473777704768713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,2048,0.006934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,2560,0.015829333000712924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,4096,0.025949334104855854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,1024,0.010712888505723743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,7168,0.07656888829337226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,256,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,3584,0.00795199970404307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,2560,0.013898667362001208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,12288,0.04140533341301812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,12288,0.022865777214368183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,6144,0.02916533417171902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,16384,0.022617777188618977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,5120,0.009748444788985783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,3584,0.011039111349317761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,256,0.007607999775144789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,2560,0.030319998661677044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,2560,0.01310755560795466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,5120,0.056835558679368764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,3072,0.011471110913488599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,512,0.01682844426896837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,1024,0.006235555642180973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,2560,0.013751111096805997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,5120,0.14620178275638157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,4096,0.007685333490371704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,256,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,12288,0.061624891228146024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,1536,0.024244444237815008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,32,0.002193777718477779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,16384,0.1419102218416002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,3584,0.034966223769717746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,1024,0.009913778139485253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,12288,0.00924800005224016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,32,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,10240,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,3072,0.007240888973077138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,128,0.01475288967291514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,512,0.02089244458410475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,768,0.0027306667632526825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,256,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,768,0.002826666666401757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,128,0.01346577786737018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,128,65536,0.07733155621422662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,5120,0.07310666640599568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,128,0.016159999701711867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,1024,0.003152888889114062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,256,0.0038497779104444715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,3072,0.08962133195665146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,10240,0.03094666533999973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,1536,0.03671288821432326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,3072,0.1410693327585856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,3584,0.026887999640570745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,12288,0.051015112135145396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,32,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,2048,0.02194044490655263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,3072,0.007538666327794393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,768,0.007592000067234039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,2048,0.03708711266517639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,6144,0.009343999955389235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,32,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,65536,0.13142400317721897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,65536,0.058166219128502734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,32,0.006622222148709827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,768,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,3584,0.018858666221300762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,2048,0.007280000381999546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,5120,0.008299555215570662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,32,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,4096,0.014764444695578681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,2048,0.006903999795516332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,65536,0.16857955190870497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,5120,0.015469332536061605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,1536,0.00630044440428416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,7168,0.07616621918148465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,3072,0.013807999591032663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,768,0.002850666642189026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,768,0.00903733323017756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,256,0.01718222267097897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,16384,0.04196622305446201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,256,0.03325422273741828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,12288,0.20796799659729004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,65536,0.20104800330268013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,3584,0.008053333395057255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3072,256,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,1536,0.010760889285140567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,8192,0.013831111292044321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,2048,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,1024,0.014152889450391134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,1536,0.0035111109415690103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,64,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,3584,0.06305333640840319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,3072,0.013427555561065674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,3072,0.015829333000712924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,7168,0.013783999615245394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,2048,0.010985777609878115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,12288,0.005877333382765452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,3072,0.00833422193924586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,8192,0.017243555850452848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,6144,0.013449778159459433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,7168,0.040567109982172646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,512,0.013391999734772576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,512,0.005896889087226655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,6144,0.012375999655988483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,12288,0.014519110321998596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,12288,0.017176000608338248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,65536,0.023745778534147475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,2560,0.017120000388887193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,3072,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,2560,0.005222222043408289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,3584,0.004545777622196409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,256,0.012067555553383298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,7168,0.018558222386572096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,1024,0.012875555290116204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,2048,0.02510844502184126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,10240,0.007584888901975419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,8192,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,5120,0.040440890524122454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,7168,1.3297644721137154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,32,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,128,0.035593777894973755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,256,0.01647911138004727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,1536,0.006994666324721442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,4096,0.009874666730562845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,2048,0.028135110934575398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,256,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,6144,0.2502675586276584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,6144,0.01872977779971229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,32,0.014313777287801107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,8192,0.021632888250880774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,1024,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,3072,0.008946666287051307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,65536,0.9792630937364367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,4096,0.008719111482302347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,3584,0.01854044364558326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,32,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,32,0.00928177767329746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,5120,0.015655999382336933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,16384,0.041759110159344144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,65536,2.0141883426242404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,768,0.011056889262464313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,65536,0.02972088919745551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,2560,0.02500711050298479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,5120,0.01237777786122428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,10240,0.01867555578549703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,128,0.0041440000964535605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,6144,0.020078221956888836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,2560,0.007294221884674496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,2048,0.008062221937709385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,3584,0.025362667110231187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,16384,0.37157689200507266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,32,0.0018639999131361644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,16384,0.040639999839994646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,12288,0.9104444715711805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,3072,0.03305066625277201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,4096,0.028346667687098186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,768,0.005234666582610872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,1024,0.006759111252095964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,256,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,2048,0.17196888393825957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,768,0.006614222294754452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,768,0.005607111172543631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,3584,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,12288,0.06625244352552626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,4096,0.011173333558771344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,12288,0.0372631119357215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,256,0.015427556302812366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,2560,0.01275111072593265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,65536,7.608283148871528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,512,0.025389333566029865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,8192,0.18169155385759142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,128,0.002171555534005165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,3584,0.004924444274769889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,4096,0.010377777947319878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,2560,0.013109332985348172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,16384,0.11696888340844049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,1536,0.036631110641691424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,1536,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,10240,0.012595555848545499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,768,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,7168,0.011034666664070554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,65536,0.01683288812637329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,3072,0.02068800065252516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,65536,0.050352000527911715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,3584,0.02660177813635932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,1536,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,3584,0.023316444622145757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,32,0.005423111220200856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,5120,0.03868888815244039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,768,0.0028213332924577924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,3072,0.08804800113042195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,1024,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,2560,0.009672889278994666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,768,0.021895999709765118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,12288,0.020090666082170274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,1536,0.015136000182893542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,768,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,3584,0.008008889026112026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,1536,0.006299555715587404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,64,0.012851555314328937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,6144,0.01001422190003925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,1024,0.004543999830881755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,768,0.007221333682537079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,3584,0.4526018036736383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,8192,0.010978666444619497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,2560,0.007631999750932057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,512,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,12288,0.014454222387737699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,64,0.035939554373423256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,16384,0.0199608885579639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,4096,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,3072,0.09264444642596775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,256,0.006762666834725275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,10240,0.012519111235936483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,16384,0.009319110876984065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,65536,0.14431999789343938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,5120,0.029278222057554457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,7168,0.015105777316623263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,2560,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,2560,0.014763555592960782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,128,0.022589332527584497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,1024,0.018181333939234417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,8192,0.044108443790011935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,32,0.005974222388532426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,64,0.002833777831660377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,5120,0.034035556846194796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,7168,0.014196443888876172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,256,0.010823999842007955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,16384,0.24897776709662545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,2560,0.02557777696185642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,128,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,1536,0.006745777610275481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,768,0.03677422139379714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,65536,0.5578142272101508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,12288,0.018464000688658822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,32,0.002061333300338851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,2560,0.026865777042176988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,2560,0.00656088896923595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,3584,0.007655999726719326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,6144,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,2560,0.00490133340160052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,1024,0.006597333484225803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,512,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,32,0.001812444378932317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,6144,0.12841778331332737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,16384,0.084897776444753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,65536,0.07580000162124634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,2048,0.07137244277530246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,256,0.0053671110007498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,512,0.013053333593739403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,128,0.010447999669445885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,2048,0.006210666563775804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,128,0.012431110772821637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,256,0.005196444276306364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,2048,0.004295999805132548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,5120,0.030582222673628066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,256,0.015323556131786771
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,512,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,768,0.0032142222755485107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,3584,0.014141332772043018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,4096,0.09769066837098862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,12288,0.02348711093266805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,2048,0.01458666721979777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,1024,0.00342755557762252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,768,0.013711999687883588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,7168,0.02935289012061225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,128,0.011197333534558615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,128,0.0021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,3584,0.03626311156484816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,3072,0.007518221934636434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,256,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,2560,0.030240890052583482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,768,0.00350222239891688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,10240,0.004912888838185204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,1024,0.014474666780895658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,2048,0.0069751110341813825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,2048,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,768,0.005537777725193236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,6144,0.01819022165404426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,768,0.016131555040677387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,256,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,128,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,10240,0.3547706604003906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,6144,0.02407288882467482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,128,0.014833778142929077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,256,0.017658667431937326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,1024,0.015411555767059326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,1024,0.006074666563007567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,3072,0.019321777754359774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,3584,0.02864177690611945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,5120,0.05053244365586174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,12288,0.031111111243565876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,768,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,3584,0.006572444405820634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,4096,0.008061333662933773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,256,0.0052666668262746595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,10240,0.1605866617626614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,12288,0.04124088750945197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,768,0.008642666869693333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,3584,0.00981600003110038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,1536,0.11327732933892144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,65536,0.06300355328453912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,12288,0.019800000720553927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,32,0.0018311111877361934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,64,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,2560,0.015488889482286243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,65536,0.12548089027404785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,1024,0.005671111245950063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,3072,0.017306667235162523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,7168,0.010071111222108206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,512,0.0069013333155049225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,1536,0.014787554740905762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,32,0.0063031112982167145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,65536,0.34517955780029297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,7168,0.011396444506115384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,12288,0.0507226652569241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,10240,0.045999109745025635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,128,0.015121777852376303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,2560,0.022359111242824133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,64,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,4096,0.016162667009565566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,4096,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,2560,0.0122079998254776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,6144,0.04429689049720764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,64,0.0024951110697454875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,32,0.011024889018800525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,12288,0.0479928884241316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,7168,0.010037333601050908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,12288,0.02204977803760105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,7168,0.019892444213231403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,8192,0.015322667029168872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,128,0.002128888956374592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,3584,0.014105778601434497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,128,0.0229795558585061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,1024,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,3072,0.04244355691803826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,2048,0.00563466673096021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,5120,0.012057777908113269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,32,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,5120,0.019188443819681805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,10240,0.026224000586403742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,8192,0.02337155573897892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,8192,0.020026667250527274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,128,0.004956444518433677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,256,0.19369067086113823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,2560,0.010311110979980892
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,3072,0.006956444846259222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,2048,0.007531555162535773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,10240,0.031042665243148804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,65536,0.9325404696994357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,7168,0.10174310869640774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,16384,0.04223111271858215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,4096,0.009303111169073317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,1024,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,3072,0.030008888906902734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,32,0.004182222402758068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,5120,0.3757377730475531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,6144,0.03442400031619602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,7168,0.021295999487241108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,65536,0.0747893320189582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,5120,0.010213333699438307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,2560,0.005043555465009478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,1536,0.015560888581805758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,2560,0.01612622208065457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,128,0.01202133297920227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,3072,0.007248000138335758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,768,0.015448000695970325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,8192,0.033194667763180204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,6144,0.009594666461149851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,2560,0.05360888772540622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,7168,0.010327999790509542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,256,0.007600888609886169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,16384,0.023352889551056757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,2048,0.004286222159862518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,64,0.0052728888889153796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,64,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,3072,0.2150204446580675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,128,0.005495999836259418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,3072,0.0140666663646698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,10240,0.017335999343130324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,1536,0.018613333503405254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,16384,0.07599822017881605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,1536,0.006649777707126405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,6144,0.01679999960793389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,1024,0.020534222324689228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,256,0.014487110906177096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,256,0.015050666199790107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,4096,0.014648889501889547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,64,0.0018044445249769424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,3072,0.01828266680240631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,768,0.006008000009589725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,8192,0.011064888702498542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,8192,0.05894844399558174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,32,0.008991999758614434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,1536,0.02102222210831112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,5120,0.24766490194532606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,6144,0.010010666317409938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,8192,0.031178666485680476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,12288,0.03429777754677667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,12288,0.0348417792055342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,1536,0.023013333479563396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,3072,0.007307555940416124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,7168,0.011330666641394297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,768,0.005616000129116907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,2560,0.023578666978412207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,8192,0.01089511149459415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,3584,0.008593777815500895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,256,0.006227555374304454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,6144,0.010033778018421596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,10240,0.020266667008399963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,128,0.005244444641802046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,7168,0.03978755407863193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,4096,0.011058666639857821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,6144,0.011343110766675739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,5120,0.0075937774446275495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,1536,0.00655377780397733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,16384,0.6848293410407172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,6144,0.04862755537033081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,12288,0.038763556215498186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,128,0.006233777850866318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,65536,0.06323555443021986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,1024,0.005953777581453323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,4096,0.01718577742576599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,128,0.005444444302055571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,10240,0.015322667029168872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,3584,0.007817777494589487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,4096,0.014807111687130399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,16384,0.028732445504930284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,1024,0.008854221966531541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,1024,0.00313688897424274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,2048,0.03363022208213806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,2048,0.02329866588115692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,6144,0.03291555576854282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,128,0.0028008888992998335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,3584,0.096160888671875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,1024,0.0038311113086011675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,3072,0.006984000404675801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,12288,1536,0.015971556305885315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,256,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,4096,0.009639110830095079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,64,0.010016889207892949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,512,0.022320000661744013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,6144,0.013433777623706393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,8192,0.019521777828534443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,12288,0.02117066747612423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,128,0.013071999781661563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,7168,0.11047199699613784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,2560,0.007252444823582967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,768,0.06708088848325941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,65536,1.451525370279948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,12288,0.012090666426552666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,8192,0.27712800767686635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,16384,0.052687111828062266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,1536,0.04592622319857279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,4096,0.12352089087168376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,10240,0.0350471105840471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,12288,0.017503110898865592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,10240,0.011398221883508893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,16384,0.052961776653925575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,1536,0.010016000105275048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,5120,0.012810666528013019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,8192,0.008752889103359645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,7168,0.043860445419947304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,7168,0.358334223429362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,6144,0.012050666742854647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,768,0.003867555409669876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,768,0.006247111078765657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,512,0.01737866633468204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,5120,0.0206675562593672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,1536,0.004651555584536659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,6144,0.02297244469324748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,6144,0.08476444747712876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,128,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,128,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,65536,0.0966871115896437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,7168,0.03850311040878296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,4096,0.023074666659037273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,256,0.007009777757856581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,6144,0.014447111222479077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,65536,0.06262666649288602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,16384,0.023353776997990076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,3584,0.016169778174824186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,3584,0.021127111381954614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,3584,0.008401778009202745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,1536,0.016827555166350473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,1024,0.013869333598348828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,5120,0.016237333416938782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,12288,6144,0.053336888551712036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,16384,0.06896799802780151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,6144,0.004922666483455234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,65536,0.0675813357035319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,256,0.01714666684468587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,2048,0.006831999868154526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,2048,0.009768000079525841
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,4096,0.01259022206068039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,1024,0.018868444694413077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,5120,0.013623110949993134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,3584,0.024199111594094172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,256,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,2048,0.021290666527218286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,768,0.006213333457708359
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,4096,0.01108000013563368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,12288,0.03745955559942458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,2048,0.009323555562231276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,1536,0.011163555913501315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,768,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,128,0.008301333420806462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,10240,0.011733333269755045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,3072,0.012712000144852532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,4096,0.0781244436899821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,12288,0.02647022240691715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,3584,0.06469511323504977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,65536,0.5531635814242893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,6144,0.010028444230556488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,7168,0.025803556044896443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,2560,0.014470222923490735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,1024,0.02469066613250309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,12288,0.050049778487947255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,65536,0.28355643484327525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,3072,0.012111110819710625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,2048,0.013774221969975365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,5120,0.009849777652157677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,6144,0.013740444348918067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,4096,0.007956444389290279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,256,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,2048,0.017436444759368896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,64,0.005205333232879639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,3072,0.00424799985355801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,3584,0.12176888518863255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,10240,0.039938665098614164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,4096,0.016393777396943834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,2048,0.010323555933104621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,128,0.0018151110659042995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,10240,0.008823999928103553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,65536,0.08847822083367242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,5120,0.009326222042242685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,2048,0.011415999796655444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,512,0.0057813334796163775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,512,0.009692444569534725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,2048,0.030559112628300984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,7168,0.01036977767944336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,3584,0.048997332652409874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,5120,0.017246221502621967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,512,0.10955378082063462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,10240,0.025232000483406916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,256,0.006592000110281839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,2560,0.00794933322403166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,4096,0.017551110850440133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,768,0.016918222109476726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,8192,0.010439111126793755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,65536,0.06007555458280775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,512,6144,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,3072,0.029138667715920344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,8192,0.0302746660179562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,6144,0.01034044474363327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,256,0.0021644443687465456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,5120,0.009874666730562845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,512,0.002871111035346985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,16384,0.04125244418780009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,10240,0.016177778442700703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,10240,0.08838488658269246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,5120,0.015296889675988091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,3072,0.020287111401557922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,768,0.005639111002286275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,7168,0.029416001505321924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,5120,0.034197333786222674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,12288,0.226511107550727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,32,0.001812444378932317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,256,0.002528888897763358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,2048,0.006567111031876669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,7168,0.023355555203225877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,65536,0.203073779741923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,8192,0.04638933473163181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,1536,0.006809777683681912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,2048,0.014771555860837301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,256,0.026920000712076824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,32,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,1536,0.006236444330877728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,5120,0.011387555963463254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,64,0.005200888961553574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,5120,0.016876444220542908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,1536,0.0038604442444112566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,2560,0.013809777796268463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,2048,0.006940444310506185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,256,0.014493332968817817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,5120,0.014463999205165438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,10240,0.07646933529112074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,8192,0.017643555998802185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,32,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,12288,0.013750221994188098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,128,0.005324444423119227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,3072,0.012067555553383298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,1536,0.007648888561460707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,7168,0.026072000463803608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,256,0.007984889050324758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,16384,0.018553778529167175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,256,0.0024622221373849446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,128,0.004879111217127906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,10240,0.014130666851997375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,4096,0.02772355576356252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1536,65536,0.036308444208569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,8192,0.014759999182489185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,7168,0.009648888475365108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,2560,0.006577777779764599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,128,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,512,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,16384,0.015598222613334656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,2560,0.031786666976081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,3584,0.00796977761718962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,5120,0.008946666287051307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,512,0.011015111373530494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,4096,0.0079644446571668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,3072,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,7168,0.2853235668606228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,768,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,256,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,7168,0.015311110350820752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,6144,128,0.002830222249031067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,10240,0.18295644389258492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,10240,0.017463111215167575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,2048,0.006578666468461354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,5120,0.011659555964999728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,7168,0.013653332988421122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,512,0.005101333475775189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,10240,0.012075555821259817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,2048,0.004498666773239772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,3072,0.00960711141427358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,5120,0.014927110738224454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,7168,0.03595911131964789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,65536,0.06318044662475586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,64,0.002193777718477779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,6144,0.00887822194231881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,7168,0.024999111890792847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,128,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,1024,0.019098665979173448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,7168,0.014477334088749357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,6144,0.23120265536838105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,1536,0.004919111314747068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,16384,0.1679262187745836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1536,10240,0.012070222033394707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,1536,0.01974933346112569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,7168,0.014760888285107084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,10240,0.042725334564844765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,8192,0.026386666629049513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,32,0.005352889084153705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,3072,0.04948000113169352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,768,0.0032311110860771606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,3584,0.0075786668393347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,128,0.005260444349712796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,12288,9.48626963297526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,6144,0.03753777676158481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,6144,0.028488000233968098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,1024,0.006120000034570694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,8192,0.023952888117896184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,16384,0.022799111074871484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,12288,0.015052444405025907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,3072,0.010767111347781287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,10240,0.057814220587412514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,2560,0.01144266708029641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,1536,0.5416426658630371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,32,0.003499555504984326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,2560,0.014085332552591959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,7168,0.15015378263261583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,1024,0.004559111264016894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,5120,0.015508444772826301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,4096,0.02286400066481696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,5120,512,0.005902222047249476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,1024,0.028976887464523315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,1536,0.006342222293217977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,7168,0.01198755535814497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,1024,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,12288,0.020789333515697055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,6144,0.04712266723314921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,128,0.20911999543507895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,2560,0.012510221865442065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,6144,64,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,512,0.05324800146950615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,16384,0.020908445119857788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,2560,0.012992888689041138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,2048,0.0041946665280395085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,768,0.009444444543785518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,1536,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,768,0.005843555761708154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,1024,0.006015111174848344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,128,0.014465777410401238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,1536,0.01313866674900055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,4096,0.008328888979223039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,256,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,16384,0.020630222227838304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,6144,0.009570666485362584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,16384,0.027760000692473516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,64,0.0018257778137922287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,12288,0.029986666308508977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,64,0.002444444431198968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,6144,0.008657777474986183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,1536,0.003875555677546395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,2048,0.004304000073009067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,128,0.0021751111166344746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,512,0.0052826665341854095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,512,0.019169777631759644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,3584,0.008303110798199972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,65536,0.11966310607062446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,6144,0.016203555795881484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,4096,0.00795999997191959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,65536,0.02567822237809499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,1024,0.006085333310895496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,768,0.006773333168692059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,6144,0.009641778137948778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,3072,0.014767999450365702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,6144,0.013055110971132914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,16384,0.06488177511427137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,2048,0.007215110792054071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,2560,0.08928889036178589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,3072,0.019582221905390423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,6144,0.08424888716803657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,128,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,128,0.0019466665883858998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1536,2560,0.013984888792037964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,128,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,3072,0.01406311160988278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,4096,0.0963395569059584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,256,0.029454221328099568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,3584,0.017367111312018502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,128,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,1024,0.003921777837806278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,128,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,2560,0.007282666862010956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,7168,0.01723377737734053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,12288,0.010015111002657149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,12288,0.014632888966136508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,8192,0.01069866700304879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,1536,0.01854933301607768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,768,0.014458666245142618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,16384,0.02000711030430264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,65536,0.24800001250372994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,3584,0.019559999306996662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,5120,0.03013333347108629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,3072,0.06990489032533434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,16384,0.7333217726813422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,6144,0.028196444114049275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,256,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,3072,0.007263999846246507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,6144,0.024357333779335022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,5120,0.030582222673628066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,1536,0.014192889134089151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,5120,0.0738604466120402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,512,0.0031617778456873367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,128,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,768,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,3584,0.008895999855465358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,2048,0.011720000041855706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,65536,0.061653335889180504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,5120,0.039136889908048846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,1024,0.0042053332759274375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,32,0.007603555917739868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,512,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,1536,0.02587022218439314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,64,0.0022008888837363985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,12288,0.04283200038803948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,256,0.015127110812399121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,1536,0.006250666661394968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,5120,0.012356444365448423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,6144,0.13702132966783312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,12288,0.022631999519136217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,128,0.005973333285914526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,1536,0.006305777778228124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,1536,0.004224888980388641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,2560,0.020967110991477966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,512,0.006844444407357111
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,3584,0.010120000276300643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,1024,0.009667555491129557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,768,0.01650399963061015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,65536,0.06817511055204603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,512,0.00905600024594201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,3072,0.02088266611099243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,1024,0.004527111020353106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,64,0.002141333288616604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,12288,0.02032533288002014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,3072,0.04455555478731791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,2048,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,128,0.009945777555306753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,12288,0.02770400047302246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,7168,0.0185217774576611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,5120,0.011665778027640449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,12288,0.025812443759706285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,12288,0.018096889058748882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,3072,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,2048,0.022223111655977037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,8192,0.1433653301662869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,768,0.005211555709441503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,128,0.1015804476208157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,256,0.015834665960735746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,8192,0.022080888350804646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,2560,0.029410666889614526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,5120,0.015297777122921415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,1024,0.007607111500369177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,64,0.006635555376609166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,256,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,2560,0.8792871369255914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,4096,0.011683555940786997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,10240,0.006614222294754452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,2560,0.007990222010347579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,768,0.017493334081437852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,128,0.033872889147864446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,2048,0.006811555474996567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,128,0.010031110710567897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,4096,0.01646844380431705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,16384,0.018961777289708454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,4096,0.011925333076053195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,7168,0.09118933147854275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,10240,0.05962933434380425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,10240,0.02253688871860504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,65536,3.8791546291775174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,3072,0.0387644436624315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,12288,0.04058400127622817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,3584,0.006579555571079254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,128,0.005723555468850666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,7168,0.5795875655280219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,12288,0.11948800086975098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,5120,0.03808000021510654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,2560,0.011424889167149862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,5120,16384,0.08505244387520684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,10240,2.1053892771402993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,5120,0.025776000486479864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,256,0.01054044481780794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,2560,0.05889422363705105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,1536,0.02868177824550205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,1536,0.05154488815201653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,6144,0.015824000040690105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,2048,0.03504622313711379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,768,0.00904444439543618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,1024,0.018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,256,0.019094222121768527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,3072,0.048771556880739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,4096,0.0778942240609063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,128,0.0028231110837724474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,1024,0.01070844464831882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,2560,0.05167733298407661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,12288,0.08915911118189494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,16384,128,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,512,0.005964444329341252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,7168,0.0841368900405036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,16384,0.0322106679280599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,3584,0.015607999430762397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,1024,0.01475644442770216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,768,0.01683022247420417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,7168,0.08048622475730048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,2048,0.015814221567577787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,2560,0.007256000406212277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,16384,0.044666667779286705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,8192,0.02807733416557312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,12288,0.02297244469324748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,2048,0.009347555538018545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,1536,0.01681777834892273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,4096,0.017895110779338412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,16384,0.021521778570281133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,8192,0.039423998859193586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,12288,0.013778666655222574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,4096,0.1087964442041185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,5120,0.030277334981494482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,128,0.010375111467308467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,8192,0.011018666956159802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,10240,0.046181334389580615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,768,0.015817777978049386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,3072,0.008984000318580205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,8192,0.075382219420539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,5120,0.008325333396593729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,2560,0.02089244458410475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,8192,0.02666933337847392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,8192,0.01648622254530589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,512,0.012744888663291931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,3584,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,3584,0.03868533174196879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,2560,0.013071111506885953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,65536,2.036672804090712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,2560,0.008775999976529015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,3072,0.015861334072219003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,768,0.017122666041056316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,2048,0.017863111363516915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,3584,0.01440355512830946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,12288,0.4697102440728082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,6144,0.051606221331490405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,32,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,256,0.002483555633160803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,7168,0.013824000126785703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,10240,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,64,0.0022400000856982335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,512,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,1024,0.006240889016124938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,6144,0.024757333927684363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,512,0.017510222064124215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,2560,0.009528888596428765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,3072,0.03682400120629205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,10240,0.014454222387737699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,4096,0.010314666562610203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,8192,0.038373334540261164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,4096,0.008609777523411645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,7168,0.019401777121755812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,5120,0.09927999973297119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,3584,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,1024,0.01906133360332913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,7168,0.03986488779385885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,256,0.0052871112194326185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,128,0.006948444578382704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,16384,0.07369955380757649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,256,0.03496444556448195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,768,0.009306666751702627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,1024,0.006280000011126201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,2560,0.02015466656949785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,8192,0.021888888544506494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,128,0.005554666535721884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,2048,0.006506666541099548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,512,0.01578044394652049
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,65536,0.1330782175064087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,1536,0.006093333164850871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,5120,0.03707999984423319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,1536,0.006638222270541721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,4096,0.07356711228688557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,4096,0.011720889144473605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,1024,0.006217777729034424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,6144,0.019558222757445443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,10240,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,10240,0.025681777132882014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,6144,0.014879110786649915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,10240,0.01734133395883772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,128,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,3584,0.007595555649863348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,6144,0.02711733347839779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,768,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,4096,0.019959110352728102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,2560,0.012381333443853589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,8192,0.020407110452651978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,7168,0.017832888497246634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,12288,0.4096088939242893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,256,0.017457778255144756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,256,0.007583111524581909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,16384,10240,0.052982221047083534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,10240,0.0343848897351159
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,12288,0.0359208881855011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,1024,0.00792266676823298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,64,0.009992889232105678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,3072,0.040553778409957886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,1024,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,8192,0.010426667001512317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,5120,0.10052977667914496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,12288,0.04286400145954556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,2560,0.014122666584120857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,6144,0.06476444668240018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,16384,0.03328533305062188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,65536,0.10336177878909641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,1536,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,256,0.002505777817633417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,1536,0.0035297779573334586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,1024,0.019832000136375427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,7168,0.004893333133724001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,64,0.04658755660057068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,16384,0.03264800045225356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,1024,0.08028711213005914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,768,2560,0.007153777612580194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,768,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,16384,0.016840888394249808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,5120,0.012703999876976013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,65536,0.009133333133326637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,64,0.0018320000833935207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,16384,0.042378667328092784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,1536,0.006239111224810283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,256,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,16384,0.09899733463923137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,12288,0.11045244667265151
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,1536,0.00998488896422916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,2560,0.03349688980314467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,10240,0.01482488877243466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,16384,0.01664533383316464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,12288,0.06355644596947564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,2048,0.004536889079544279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,512,0.0058959999846087555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,512,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,16384,0.03295200069745382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,12288,0.01689511040846507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,32,0.04766844378577339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,512,0.04264711009131538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,512,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,10240,0.015415110521846347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,3584,0.014448000325096978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,3072,0.004899555610285865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,4096,0.007670222057236566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,3584,0.008345777789751688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,512,0.013464888764752282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,10240,0.1180000040266249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,128,0.0031866667171319327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,65536,0.06265244219038221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,6144,0.019889778561062284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,1024,0.0031946665710873077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,7168,0.011355555719799466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,7168,0.011177777416176267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,10240,0.012111110819710625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,1536,0.017250667015711468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,128,0.010010666317409938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,4096,0.11443555355072021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,6144,0.014736889137162102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,3072,0.004889777965015835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,2048,0.007594666547245449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,7168,0.0278479986720615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3584,128,0.005227555417352253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,16384,0.007567110988828871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,5120,0.009683555199040307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,256,0.0021582223061058256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,12288,0.08690222104390462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,4096,0.00612088872326745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,4096,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,3584,0.007090667055712805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,8192,0.17014755143059623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,7168,0.011149333582984077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,1024,0.005611555443869696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,10240,0.005628444254398346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,1024,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,16384,0.06904621918996175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,768,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,10240,0.1977271106508043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,1024,0.01719466679626041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,65536,0.12580177519056532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,16384,0.03908088803291321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,1024,0.00564533347884814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,768,0.009005332986513773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,7168,0.03677333394686381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,768,0.01549600064754486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,256,0.017214222086800467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,7168,0.013726222018400827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,512,0.008287111090289222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,256,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,3072,0.013108444710572561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,7168,0.011236444115638733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,2560,0.010443555812040964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,2048,0.015321777926550971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,16384,0.037179556157853864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,12288,0.03074755602412754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,3072,0.009704888694816166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,4096,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,512,0.005578666511509154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,65536,3.561912112765842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,16384,0.13973244031270346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,65536,0.09474133120642768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,3072,0.0070257774657673305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,16384,0.028160889943440754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,3072,0.019842666056421068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,3584,0.04497866829236349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,256,0.006606222026877933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,3072,0.006924444602595435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1024,768,0.018201778332392376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,4096,0.007982222570313348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,2048,0.016862221890025668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,12288,0.01644711030854119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,256,0.005384888913896348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,5120,0.014708444476127625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,5120,0.03252622154023912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,6144,0.014102222190962898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,8192,0.019119110372331407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,65536,0.07079378101560804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,8192,0.01070844464831882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,128,0.014073777529928418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,3072,0.031368000639809504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,16384,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,10240,0.05241511265436808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,12288,0.02997422218322754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,6144,0.032080888748168945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,2048,0.006222222414281633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,2048,0.08416089084413315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,2560,0.029952890343136255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,512,0.005579555614127054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,6144,256,0.005910222315125995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,128,0.004915555732117759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,7168,0.009655999640623728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,256,0.010839111275143094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,3072,0.03690222236845229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,512,0.00444088876247406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,10240,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,768,0.0046471113132105935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,6144,6144,0.03638311227162679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,12288,0.0871466663148668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,7168,0.01239733315176434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,3072,0.017162667380438913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,3584,0.09128710958692764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,3584,0.008289777570300633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,256,0.0021955555097924336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,256,0.01681866745154063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,256,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,3584,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,2560,0.010021333065297868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,2048,0.015478221906556023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,8192,0.01571466690964169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,16384,0.02107911143038008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,8192,0.03695822093221877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,8192,0.040735999743143715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,6144,0.1988666719860501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,512,0.005375999957323074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,3584,0.011724444727102915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,256,0.0024906665914588505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2560,128,0.0059164443777667145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,8192,0.03366488880581326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,3584,0.06291288799709745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,4096,0.026549332671695288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,6144,0.01034755590889189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,3584,0.033837334977255926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,65536,0.04946666624810961
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,1536,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,256,0.0021119999388853707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,3584,0.019732443822754752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,8192,0.019293333093325298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,65536,0.2573617829216851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,256,768,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,7168,0.013435555828942193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,5120,0.014967999524540372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,6144,0.08972266647550796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,512,0.0033155555526415506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,768,0.005639111002286275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,512,0.006924444602595435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3584,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,3072,0.013433777623706393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,10240,7168,0.03350844317012363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,8192,0.03186310993300544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,6144,0.012553777959611682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,65536,0.07145600186453925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,256,0.017047999633683097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,256,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,2560,0.006511111226346757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,5120,0.018941332896550495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,3584,0.028839111328125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,64,0.013764444324705334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,1024,0.008190222084522247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,3584,0.008018666671382057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,1536,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,6144,0.01921955578856998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,4096,0.007633777956167857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,10240,0.11247200436062282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,8192,0.028604444530275133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,16384,1.0800817277696397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,4096,0.025055999557177227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,8192,0.022964444425370958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,128,0.0018817778262827133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,256,0.0076142218377855085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,10240,0.012056888805495368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,128,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,7168,0.02195466723706987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,2048,0.007080888582600488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,6144,0.021648888786633808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,2048,0.018574222922325134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,32,0.005583999885453119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,512,0.005938666562239329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,4096,0.01477866702609592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,512,0.002509333400262727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,10240,0.05143910977575514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,8192,0.015448888142903646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,10240,4096,0.026168000366952684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,1024,0.017330666383107502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,7168,0.01575644479857551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,32,0.007924444145626491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,6144,0.009039999710188972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,8192,0.021645334031846788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,8192,0.011363555159833698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,64,0.001783999924858411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,4096,0.00795288880666097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,7168,0.004859555512666702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,8192,0.028387556473414104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,3072,0.004232888834344016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,16384,0.022413333257039387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,768,0.007284444239404466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,7168,0.011501332951916588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,7168,0.01960000064637926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,1024,0.0123831108212471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,2048,0.006637333167923822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,128,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,8192,0.02203111184967889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,128,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,12288,0.013944889108339945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,512,0.008634666601816813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,128,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,4096,0.007976888782448238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,768,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,768,0.011728888584507836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,32,0.005241777747869492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,1536,0.06788444519042969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,8192,0.0103511114915212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,512,0.002520888836847411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,7168,0.08212000131607056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,8192,0.039678222603268094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,5120,0.09565955400466919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,32,0.007814221911960179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,512,0.0028764444092909494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,10240,0.05522933271196154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,16384,16384,0.09140444464153713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,65536,0.3314533233642578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,6144,0.03713688916630215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,4096,0.04789066645834181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,12288,0.007313778003056844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,768,0.007122666471534305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,768,0.006671999891599019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,6144,0.008818666968080733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,7168,0.13261600335439047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,3584,0.05552711089452108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,12288,0.02860800094074673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,2560,0.0595031115743849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,12288,0.005410666680998272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,12288,0.13380532794528538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,6144,0.1393644412358602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,768,0.011211555865075854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,7168,0.02405066622628106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,10240,0.01590044465329912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,5120,0.016472889317406546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,128,0.0058062221441004015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,7168,0.016211556063758004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,2048,0.041872001356548734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,1536,0.00423555572827657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,1536,0.017866666118303936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,12288,0.018232888645595975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,65536,0.13599467277526855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,256,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,2048,0.011142222417725457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,2560,0.02824977702564663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,2048,0.0418888893392351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,12288,256,0.006265777680608962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,8192,0.04212622178925408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,7168,0.09365955326292251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,65536,0.031239999665154353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,8192,0.030947556098302204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,256,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,65536,0.23867466714647081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,768,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,4096,0.0076453329788313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,1024,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,7168,7168,0.02070488863521152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,2048,0.007274666594134436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,65536,0.27242223421732586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,3584,0.01792622274822659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,128,0.021007110675175984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,1536,0.0063031112982167145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,2048,0.020827554994159274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1536,12288,0.04312622216012743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,3584,0.030949334303538006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,8192,0.03160533308982849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,6144,0.009023110899660323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1536,8192,0.02568977740075853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,2048,0.015512888630231222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,16384,0.929832034640842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,3584,0.011160888605647616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,12288,0.016746666696336534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,4096,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,3072,0.007249777515729268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,2048,0.01753866672515869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,32,0.0017928888814316855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,2048,0.008877333667543199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,3584,0.012261333564917246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,65536,0.13258666462368435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,768,0.016763556334707472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,2560,0.004851555658711327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,7168,0.08129955662621392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,3584,0.012658666405412885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,128,6144,0.016551110479566786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,16384,0.016521778371598985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,6144,0.0499875545501709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,1536,0.007716444631417592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,3072,0.012229333321253458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,16384,0.03627289003796048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1536,4096,0.009339555270142024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,512,0.020975111259354487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,6144,0.004279110994603899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,7168,0.706274668375651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,512,0.015839111473825242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,4096,0.027393778165181477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,16384,0.0878480010562473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,2560,0.0096124443742964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,5120,0.012099555797047086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,10240,0.040553778409957886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,12288,0.022982221510675218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,256,0.016156444946924847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,4096,0.010016889207892949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,6144,0.18872355090247261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,2048,0.015765332513385348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,6144,0.009675555759006077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,65536,0.07120888763003878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,1536,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,1024,0.016590222716331482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,5120,0.022241777843899194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,7168,0.016424889365832012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,512,0.00657066661450598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,5120,0.023738667368888855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,3584,0.007627555893527136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,3072,0.026181333594852026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,12288,0.03426933288574219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,10240,0.02977422210905287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,128,0.010724444356229572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,5120,0.021663111117151048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,7168,0.011676444775528379
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,2560,0.007622222105662028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,64,0.008023999631404877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,2560,0.016303999556435477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,16384,0.04059022333886888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,512,0.013423999978436364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,3584,0.00921688891119427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,5120,0.01717688971095615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,128,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,7168,0.024364444944593642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,7168,0.032984889215893216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,128,0.0052933332820733385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,5120,0.024734222226672705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,12288,0.04763466782040066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,8192,2560,0.008980444735950893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,5120,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,5120,0.058712886439429395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,8192,0.018578666779730055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,8192,0.016591111818949383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,1024,0.00282577777074443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,3072,0.007797333101431529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,4096,0.17631644672817656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,512,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,1024,0.0028880000528362063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,6144,0.18774133258395723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,12288,0.017826666434605915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,768,0.06185599830415514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,5120,0.00942488925324546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3584,512,0.014479999740918478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,12288,0.07310755385292901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,128,0.002471111093958219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,8192,0.09097955624262492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,16384,0.02706577711635166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,256,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,7168,0.019194665882322524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,10240,0.11339022053612603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,8192,0.03609688745604621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,6144,0.04849333233303494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,1024,0.016186666157510545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,10240,0.023423110445340473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,128,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,12288,1.0698578092787001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,2048,0.013412444127930535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,10240,0.03450222147835626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,512,0.02910755409134759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,128,0.00832177781396442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,128,0.00526577772365676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,65536,0.08714489142100017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,1536,0.012602667013804117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,3072,0.007551111280918121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,65536,0.12644710805681017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,16384,0.02049955560101403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,1536,0.01184177812602785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,128,0.0017946666727463405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,16384,1024,0.02698488864633772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,512,0.0022702221241262224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,64,0.00553688903649648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,2048,0.1881937715742323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,5120,0.009317333499590555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,7168,0.013459555804729462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,3584,0.0402693317996131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,3584,0.00720000018676122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,3072,0.026341333985328674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,1024,0.013793778088357715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,32,0.00666844430896971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,12288,0.06238044632805718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,2048,0.022681777675946552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,3072,0.019654222660594516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,256,0.016504888733228047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,32,0.0031582222630580268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,768,0.026157332791222468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,6144,0.06587111287646823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,10240,0.014797333214018079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,12288,0.009844444692134857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,2560,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,12288,0.0162791113058726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,128,0.002177777803606457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,10240,0.013437333206335703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,65536,0.02815199891726176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,1024,0.017172444197866652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,512,0.007986666427718269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,1536,0.011750222080283694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1024,7168,0.08636444144778782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,512,0.0028657778683635923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,6144,0.0123857781291008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,768,0.022509333160188463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,12288,0.01735377808411916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,2560,0.00553955551650789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3072,128,0.014762666490342883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,1024,0.005978666659858491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,12288,0.013733333183659447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,3584,0.02052622205681271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,768,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,3072,0.007224000162548489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,3072,0.007930667036109501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,7168,0.007015110717879401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,2048,0.013393777940008374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,16384,0.018942221999168396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,4096,0.024704888463020325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,512,0.015807999504937064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,8192,0.016520000166363187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,4096,0.010863111250930362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,2048,0.029278222057554457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,2048,0.005921777751710679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,512,0.007260444263617198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,16384,0.020738666256268818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,256,0.011683555940786997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1536,1024,0.00314666661951277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,2560,0.011520889070298938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,2560,0.020790222618314955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,5120,0.008268444074524773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,6144,0.017856000198258292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,1024,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,64,0.007848888635635376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,4096,0.007920888562997183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,10240,0.0147706667582194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,256,0.0052728888889153796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,3072,0.007237333390447829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,128,0.002128888956374592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,1024,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,8192,0.80495908525255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,1024,0.007651555869314406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,3072,0.007653333246707916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,3072,0.013206221991115145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,6144,0.05172444383303324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,3072,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,3584,0.007354666789372762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,5120,0.022915555371178523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,7168,0.010662222074137794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,768,0.007594666547245449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,768,0.0059573331640826324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,128,0.0029102222373088202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,64,0.002509333400262727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,128,0.006912888752089606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,5120,0.027436445156733196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,128,0.002157333410448498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,512,0.013746666411558786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,5120,0.06007022327846951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,7168,0.09180355734295315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,12288,0.018191999859280057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,3072,0.015203555425008139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,128,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,256,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,2560,0.006924444602595435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,65536,0.03254577848646376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,2048,0.0374293327331543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,32,12288,0.007291555404663086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,3584,0.033161779244740806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,65536,0.29858043458726674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,8192,0.012098666694429187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,2560,0.007655999726719326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,3072,0.008616000413894653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,512,0.01621866722901662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,768,0.005640000104904175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,16384,0.017208000024159748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,8192,0.011693333586057028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,3072,0.004570666700601578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,2560,0.041441778341929116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,256,0.013651555611027611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,128,0.011050666371981302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,768,10240,0.01275644451379776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,2048,0.3472462230258518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,256,0.015792000624868605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,64,0.008959111240175035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,12288,5120,0.21543733278910318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,16384,0.05190666516621908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,7168,256,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,8192,0.01073688848151101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,65536,32,0.007690666450394525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,512,0.005919999960396025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,1536,0.004548444516128964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,128,0.005081777771313985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,256,0.006627555522653792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,768,5120,0.021307556165589228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,1024,0.01741333305835724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,7168,0.026501332720120747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,32,0.0018613334331247543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,5120,1.1084533267550998
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,768,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,2048,0.007719111111429002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,65536,0.11933067109849717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,256,0.0032026666320032547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,10240,0.01791111131509145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,1536,0.003871110992299186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,3072,0.08985422054926555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,32,0.025376889440748427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,1536,0.006607111129495833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,256,256,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,512,0.014800889624489678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,2048,0.017903111047214933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,768,0.005967111223273807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,1024,0.0031831111345026228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,12288,0.03330577744377984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,2048,0.005535999933878581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,768,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,8192,0.032425777779685125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,5120,0.021219554874632094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,3584,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,1024,0.01994222237004174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,12288,0.03676355547375149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,12288,0.036130666732788086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,1536,0.006314666734801398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,64,0.006311111152172089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,5120,0.008856000171767341
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,1024,0.010049777726332346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,512,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,3584,0.007669332954618666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,8192,0.048820445934931435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,5120,0.028026666906144884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,65536,0.18911911381615532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,65536,16384,0.2965262201097277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,3584,0.016394666499561734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,1024,0.023214222656355962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,5120,0.03308622373474969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,7168,0.32294045554267037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,4096,0.028821332587136164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,2560,0.008016000191370646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,6144,1536,0.015004444453451367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,8192,0.056699554125467934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,4096,0.012039999994966718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,6144,1536,0.006611555400821898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,3072,0.03190844588809543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,512,0.002500444443689452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,128,0.0024622221373849446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,12288,0.045540445380740695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,65536,0.007177777588367462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,4096,0.008316444853941599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,4096,0.12007110648685032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,1024,0.01962222158908844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,128,0.0032675556010670135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,10240,0.010327999790509542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,768,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,2048,0.007625777688291337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,65536,0.2514248953925239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,256,0.005892444401979446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,512,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,128,0.01756977703836229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,12288,0.01478666729397244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,65536,0.05987733602523804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,3584,0.03775644302368164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,12288,0.2692755593193902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,128,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,12288,0.01475822263293796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,4096,0.02086755633354187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,6144,0.09210844172371759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,10240,0.03876711262596978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,512,0.017267554998397827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,2048,0.02625333269437154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,64,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,7168,128,0.009318222602208456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,1536,0.008563555777072906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,7168,0.05848089191648695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,1024,0.0939377744992574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,32,0.007662222617202335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,2560,0.00488711107108328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,3072,0.032891554964913264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1024,3584,0.00759733302725686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,10240,0.012143111063374413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,256,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,5120,0.020963556236690942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,2560,1.9305689069959853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,1024,0.005917333480384614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,128,0.003519111209445529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,2048,0.028194665908813477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,3072,0.01737511157989502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,128,0.006351111249791251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,12288,0.06675999694400363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,1536,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,7168,0.023288889063729182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,2560,0.043655110730065234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,3072,0.01716444392999013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,3584,0.00756888919406467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,5120,0.009030222064918941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,6144,0.013429332938459186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,4096,0.007637333538797166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,2048,0.009693333672152625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,1024,0.01613333324591319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,5120,0.011891555454995898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,768,0.005920889063013925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,3072,0.009537777966923183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,3584,0.007668444679843054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,1536,0.006587555425034628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,1536,0.026549332671695288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,2560,0.007178666690985362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,768,0.45676088333129883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,2560,0.0069822221994400024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,10240,0.03148355417781406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,1536,0.05227111114395989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,16384,0.059567113717397056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,65536,0.14486755265129939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,512,0.006640888750553131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,65536,0.06671555836995442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,256,0.0038968887594011095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,1536,0.016827555166350473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,7168,0.026639110512203638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,65536,0.09359644518958198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,1536,0.016173332929611206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,1536,0.030759112702475652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,8192,0.034522665871514216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,4096,0.008711111214425828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,4096,0.03460799985461765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,8192,0.015013333823945789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,12288,0.024285333024130926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,2048,0.006276444428496891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,4096,0.030238221089045208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,64,0.0021786666992637846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,10240,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,65536,0.14996000130971274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,3584,0.011335111326641507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,32,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,64,0.005232000102599462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,768,0.005369777894682354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,128,0.007644444704055786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,5120,0.03122666809293959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,8192,0.03071199854214986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,128,0.012105777859687805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,10240,0.016574222180578444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,4096,0.11684177981482612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,5120,0.06407911247677274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,256,64,0.0018320000833935207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,2048,0.006911110960774952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,16384,0.01795644395881229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,768,0.0320995549360911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,3072,0.019946666227446664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,2560,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,2048,0.004501333253251182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,64,0.0053342220683892565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,3584,0.04996088809437222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,3072,0.007317333585686154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,2048,0.025237333443429735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,16384,0.016915554801623028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,512,0.019671110643280875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,5120,0.012084444363911947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,2048,0.006621333460013072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,2560,0.15197600258721247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,16384,0.17499377992418078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,16384,0.01791733337773217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,6144,0.06951466533872816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,256,0.0022071111533376905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,256,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,3584,0.006913777854707506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,10240,0.019708444674809773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,1024,0.007332444190979004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,5120,0.009327111144860586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,2560,0.03423555692036947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,4096,0.06367022461361356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,1536,0.009813333551088968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,128,0.002013333348764314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,512,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,1024,0.017502221796247695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,6144,0.012992888689041138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,1536,0.008671999805503422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,3584,0.01631644368171692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,6144,0.01772533357143402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,3072,0.00452977791428566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,8192,0.3661315441131592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,65536,0.1217306719885932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,64,0.00314666661951277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,128,0.01108000013563368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,256,0.0021280000607172647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,1536,0.010679110884666443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,32,0.005261333452330695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,8192,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,512,0.0052577778697013855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,2560,0.028864890336990356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,7168,0.050864888562096484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,3584,0.008923555413881937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,12288,0.016573333077960543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,2560,0.01864977843231625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,32,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,10240,0.017991110682487488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,2048,0.0713519983821445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,5120,0.03227555420663621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,1536,0.006230222268237009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,65536,0.14727289146847194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,5120,0.018552889426549275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,1024,0.023016888234350417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,3584,0.00833511104186376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,6144,0.028130667077170476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,2048,0.01000088867213991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,128,0.006218666831652324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2048,10240,0.028488887680901423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,3584,0.004514666481150521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,8192,0.043214221795399986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,5120,0.017328000730938382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,3072,0.037343998750050865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,1536,0.01851377718978458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,768,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,256,0.015122666954994202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,128,0.0031555555760860443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,65536,0.08033688863118489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,12288,64,0.018568888306617737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,16384,0.02647022240691715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,768,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,4096,0.020232889387342665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,2048,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3584,8192,0.025245333711306255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,512,0.00684799998998642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,12288,0.030511998467975195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,2560,0.03283022178543939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,16384,0.7183661990695529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,4096,0.03720888826582167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,768,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,5120,0.014965333872371249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,1024,0.007270221908887227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,3584,0.0244151105483373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,128,0.01235466698805491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,12288,0.020561777883105807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,512,0.002855111120475663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,7168,0.0221368885702557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,12288,0.01460088955031501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,128,0.005035555611054103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,12288,0.026713778575261433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,3072,0.004690666579537922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,7168,0.014239999983045789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,256,0.0021484444538752237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,1536,0.015971556305885315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,1024,0.035912000470691256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,12288,0.01443555619981554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,512,0.0038248888320393036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,7168,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,2560,0.004534222185611725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,512,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,65536,0.08483822478188409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,64,0.017156443662113614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,2560,0.005221333354711533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,128,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,16384,0.0442791117562188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,12288,0.1280399958292643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,65536,1.1481271319919162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,3072,0.007947555846638149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,6144,0.11452799373202854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,12288,0.019308444526460435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,4096,0.04005333450105455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,2048,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,16384,0.028933333026038274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,16384,0.04555822081036038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,2048,0.005209777918126848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,4096,0.015468445089128284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,12288,0.08357066578335232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,4096,0.006626666833957036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,2560,0.008268444074524773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,7168,0.022681777675946552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,2560,0.008664888640244802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,512,0.022302221920755174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,1024,0.003863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,128,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,1536,0.018168000711335074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,32,0.003175111073586676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,2048,0.0075128889746136135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,128,0.0021564445147911706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,2048,0.020968000094095867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,512,0.014779556128713818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,1536,0.019484443797005545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,128,0.0021608887861172357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,6144,0.004607111215591431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,1024,0.05389066537221273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,1536,0.004552888787455029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2048,2048,0.019561777512232464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,65536,1.3080906338161893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,8192,0.05840088923772176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,12288,1024,0.009681777821646797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,1536,0.008027555214034187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,1536,0.008295999632941352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,1024,0.015120888749758402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,32,0.008009778128729926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,65536,2.0247609880235458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,65536,0.14134221606784397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,512,0.013597333596812354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,768,0.0028613333900769553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,1024,0.008991111483838823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,256,0.015480889214409722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,7168,0.01668177710639106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,4096,0.017595556047227647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,768,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,12288,0.31775999069213867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,256,3584,0.007345777418878343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,128,0.0842204425070021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,12288,0.0687662230597602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,256,0.017157332764731515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,256,0.0024871110088295406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,3072,0.022445332672860887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,1536,0.0053546664615472155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,256,0.01510933372709486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,5120,0.008300444318188561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,3584,0.004579555657174853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,512,0.012811555630630918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,1024,0.004565333326657613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,4096,0.008287111090289222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,5120,0.012454222473833295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,1536,0.013190222283204397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,2048,0.007456888755162557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,3584,0.09060711330837673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,32,0.013760888742076026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,768,0.01570399933391147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,32,0.0018293333964215384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,16384,0.0794080032242669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,5120,0.020092444287406076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,2560,0.004910222358173794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,3584,0.022270222504933674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,512,0.007249777515729268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,8192,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,10240,0.053005334403779775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,3072,0.006899555524190267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,12288,0.27495821317036945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,512,0.0035093331502543557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,4096,0.008010666403505538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,8192,0.08892089128494263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,12288,0.0321760012043847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,16384,0.027091556125217017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,512,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,6144,0.15124711725446913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,10240,0.2023448944091797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,768,0.012038222617573209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,5120,0.11736355887518989
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,65536,0.3662159972720676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,6144,0.012449777788586087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,5120,0.015565334094895257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,768,0.0032017777363459268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,512,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,128,0.03854044609599643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,5120,0.014132445057233175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,3584,0.00944088896115621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,768,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,7168,0.02174844510025448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,16384,0.020661332541041903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,10240,0.016864000095261466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,768,0.026512000295850966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,256,0.001834666563404931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,768,64,0.0018177777528762817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,2048,0.012313777373896705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,5120,0.02233422299226125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,4096,0.005283555636803309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,7168,1536,0.006716444260544247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,5120,0.009872888525327047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,768,0.02513155506716834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,5120,0.0391173329618242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,768,0.01750666730933719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,1024,0.014811555544535318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,6144,0.009315555294354757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,4096,6144,0.01204355557759603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,2048,0.010953777366214328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,6144,0.010712888505723743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,128,0.014782221780882942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,7168,0.04145866632461548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,2560,0.017481777403089736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,12288,2.3005875481499567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,16384,0.3900737762451172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,10240,0.023696889479955036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,1024,0.016193777322769165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,16384,0.07357422510782878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,4096,1536,0.00628977765639623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,65536,0.04638311266899109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,16384,0.07948978079689874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,3584,0.009896888501114314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,8192,0.014475555883513557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,10240,0.11354222562577988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,512,0.006212444355090459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,6144,0.02576622201336755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,5120,0.0038897775941424896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,512,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,2560,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,512,0.01612977849112617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,16384,0.016143111719025504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,3072,0.04074755642149184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,2560,0.022279111875428095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,1536,0.008623110751310984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,12288,4096,0.014362666341993542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,2560,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,5120,5120,0.031720889939202204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,2560,0.00942666663063897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,6144,0.10735644234551324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,1536,0.031599998474121094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,65536,0.05577511257595486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,3072,0.01647111111217075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,32,0.001790222194459703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,32,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,128,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,16384,0.11077244414223565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,1024,0.007609777980380588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,16384,0.05201244354248047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,4096,0.01072088877360026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,6144,0.017214222086800467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,2048,0.02219733264711168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,1536,0.008276444342401292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,7168,0.04321155614323086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,16384,0.03362222181426154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,12288,0.019899555378490023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,16384,0.025405334101782903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,768,0.13963554965125188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,2560,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,64,0.0038933331767717996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,32,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,65536,0.12096977233886719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,2560,0.013775111072593264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,5120,0.010506666368908353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,8192,0.005593777944644292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,12288,0.02498311135503981
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,2048,0.0069297779765393995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,3584,0.01036711119943195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,16384,0.033124443557527326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,1024,0.032256888018714056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,16384,2560,0.016753777861595154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,128,0.013071999781661563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,1024,0.009640889035330879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,16384,0.03728177812364366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,5120,0.027427555786238775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,16384,0.07487110959159003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1024,512,0.008911111288600499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,128,0.00556355549229516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,512,0.016950221525298226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,2048,0.005249777601824866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,7168,0.03631822268168131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,12288,0.02574933403068119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,5120,0.0111191107167138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,16384,0.014551111393504672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,256,0.016141333513789706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,128,0.0052782222628593445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,8192,0.014159111513031853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,1536,0.0042035554846127825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,65536,0.5371155738830566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,5120,0.009029332962301042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,2560,0.0791048871146308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,2560,0.013778666655222574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,6144,0.02027555637889438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,3584,0.013759999639458127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,16384,0.028799109988742407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,2048,0.05103733473353916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,256,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,3072,0.006934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,2048,0.016184000505341422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,5120,0.021595555875036452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,10240,0.010694222317801582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,512,0.015849777393870883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,128,0.015114666687117683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,7168,4096,0.012086222569147745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,4096,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,1024,0.0028435554769304064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,3584,0.03815733393033346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,65536,0.14423288239373103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,3072,0.03714399867587619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,128,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,256,0.0035413333939181436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,8192,2048,0.008607111043400234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,512,0.020285333196322124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,65536,0.1314515537685818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,1536,0.13183732827504477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,1024,0.025763556361198425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,3584,0.12526044580671522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,16384,0.025576889514923096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,128,0.0069466663731469055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,128,0.004936000125275718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,3072,0.007248000138335758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,8192,1536,0.02461244497034285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,16384,0.3670853243933783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,3072,0.015996444556448195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,12288,0.03650310966703627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,4096,0.02979288829697503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,768,0.01516533394654592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,16384,0.0251208891471227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,512,0.002495999965402815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,12288,0.009658666948477427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,32,0.0023591110689772502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,3072,0.017160000072585214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,5120,0.021572444174024794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,32,0.0021386666016446217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,64,0.0017911110901170308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,256,0.012706666356987424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,2048,0.018932445181740653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,6144,0.6528284284803603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,12288,0.03376977642377218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,1024,0.030215998490651447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,4096,0.029419554604424372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,5120,0.010008888940016428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,256,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,6144,2560,0.007466666400432587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,256,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,1536,0.007991111112965478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,768,0.005884444548024072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,2560,0.01687466601530711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,7168,0.11075022485521104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,7168,0.0247706671555837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,8192,0.038302222887674965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,6144,0.022272000710169475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,16384,0.02236888971593645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,3072,0.023414222730530634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,3584,0.01278222186697854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,2048,0.18082933955722383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,1536,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,64,0.006331555545330048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,768,0.015984889533784654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,10240,0.02264444364441766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,64,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,5120,0.008664888640244802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,10240,0.08753244744406806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,16384,0.02308711078431871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,16384,0.017549332645204332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,2048,0.4875955581665039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,3072,0.027753776974148218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,65536,0.19168266985151503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,3072,0.007897777689827813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,512,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,2048,0.006643555644485686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,4096,0.015178667174445258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2048,3072,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,65536,0.0896951092614068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,3072,0.022442667020691767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,3072,0.04079644547568427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,7168,0.01261333293384976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,256,0.019686222076416016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,3072,0.07360621955659655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,128,0.0021475555582178966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,12288,0.020265777905782063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,1536,7168,0.013065777719020844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,256,0.008280889027648503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,8192,0.3915982246398926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,7168,16384,0.9250746832953559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,3072,0.02903999884923299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,768,0.005201777650250329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,16384,0.13851822747124568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,12288,1.1774159537421334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,3584,0.00998933364947637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,3584,0.016338666280110676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,2048,0.02330400049686432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,7168,0.04195733202828301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,12288,0.024373332659403484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,10240,0.006586666736337874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,12288,0.009473777479595607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,16384,0.3209342161814372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,512,0.011828444070286222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,10240,0.1084408892525567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,3584,0.030869334936141968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,1536,0.015828443898095023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,4096,0.03626133335961236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,8192,0.01146755533085929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,65536,0.06369955672158135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,3072,0.007232888705200619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,768,0.05196088883611891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,2560,0.02164711058139801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,3072,0.013960889644092984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,7168,0.013042666845851474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,768,1024,0.013751111096805997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,7168,0.0255084451701906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,2048,0.029954665237002905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,16384,0.016474665866957772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,512,0.0024951110697454875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,768,0.019098665979173448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,8192,0.011030221978823343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,2048,0.004537777768241035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,8192,0.18301333321465385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,128,0.013790222505728403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,6144,0.03300088975164626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,7168,0.007358222372002072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,16384,0.3573191165924072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,65536,0.057130667898390025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,2048,0.0064186664919058485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,10240,12288,0.031918221049838595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,65536,7168,0.12764267126719156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3072,12288,0.027898665931489732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,512,0.0024764444679021835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,128,0.0052595556610160405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,10240,0.03130933311250474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,256,0.016534222496880423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,5120,0.009679111341635386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,3072,0.023657777243190344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,65536,0.17908799648284912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,7168,0.010035555395815108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,128,0.005195555587609609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,768,0.03010222315788269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,128,0.014488000008794995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,10240,0.03598755598068237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,4096,0.017287110288937885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,7168,0.009294222626421187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,5120,0.018797333041826885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,10240,0.06690311431884766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,5120,0.01238222254647149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,16384,0.08201777935028076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,64,0.010087110930018956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,65536,0.09195733070373535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,64,0.01179377817445331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,256,0.0021520000365045336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,4096,0.017048888736300997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,6144,8192,0.02290577855375078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,2560,0.007124444676770105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,256,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,1536,0.05579821930991279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,8192,0.20073421796162924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,4096,0.009683555199040307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,768,0.018205333087179396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,512,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,128,65536,0.013271999855836233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,3072,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,65536,0.17492622799343535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,768,0.046803557210498385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,512,0.00645955569214291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,768,0.006305777778228124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,16384,0.015848888291252982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,8192,0.01684799955950843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,12288,0.013397333522637686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,5120,0.02964622113439772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,10240,3584,0.014179555906189812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,5120,12288,0.2290924390157064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,65536,0.008986666798591614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,256,0.002487999904486868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,768,0.0026319999661710528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,1536,0.003531555748648114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,12288,0.051226665576299034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,7168,0.13054933812883165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,256,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,3584,0.004556444370084339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,16384,0.03506311111980014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,65536,0.0639431079228719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,2560,0.06155111392339071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,7168,0.05165333218044705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,32,0.001850666685236825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,3584,0.009357333183288574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,32,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,1024,0.01510933372709486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,7168,0.029394666353861492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,5120,0.010209778116808997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,7168,0.15049066808488634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,1536,0.014918223023414612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,4096,0.018535999788178336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,3072,0.0240000006225374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,10240,0.017416000366210938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,3072,0.01958044370015462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,6144,8192,0.015134221977657743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,32,0.0018586667461527719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,3072,0.007633777956167857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,7168,0.01870666609870063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,7168,0.016743110285864938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,2560,0.021759111020300124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,3072,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,128,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,1536,3072,0.023010666171709698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,5120,0.01037333326207267
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,10240,16384,0.06568621926837497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,512,0.02653955585426754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,5120,0.03288622366057502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,4096,0.020690666304694284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,4096,0.015018666783968607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,1024,0.037612444824642606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,2048,0.028656889994939167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,256,0.04077688852945963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,65536,0.06538577874501546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,2560,0.006721777634488211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,32,0.0021546667234765156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,3584,0.00778399987353219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,1536,0.050208889775806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,3072,0.004593777573770947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,6144,0.017550221747822232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,1024,0.0041848888827694785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,4096,0.009133333133326637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,256,0.015171556009186639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,512,0.005050666630268097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,2048,0.006607999818192587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,3584,0.007978666987684038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,12288,0.07165866427951388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,1536,0.008652444514963362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,64,0.002505777817633417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,32,0.04589955674277412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,8192,16384,0.04545333319240146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,4096,0.008298666940795051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,12288,0.1706906689537896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,6144,0.024941333466105994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,4096,0.034910221894582115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,1536,0.007628444168302748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,512,0.006583111153708563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,768,0.006303999986913469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,12288,0.02658044464058346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,8192,0.08967288997438218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,2048,0.0041866666740841335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,10240,0.02000799940692054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,256,0.0022524444179402455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,1536,0.007956444389290279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,1024,0.005208888815508948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,4096,0.15966755814022487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,3072,0.09722310966915554
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,5120,0.014366222752465142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,512,0.011990221838156382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,4096,768,0.016175111134847004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,16384,0.020868443780475195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,3584,0.02290133304066128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,5120,0.008622222476535374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,8192,0.009692444569534725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,768,0.0052684446175893145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,1024,0.006102222121424145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,2048,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,65536,0.11453866958618164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,4096,0.05057422320048014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,256,0.0031893334041039148
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,8192,0.01072088877360026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,128,0.012748444245921241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,128,0.002157333410448498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,3072,0.01074044406414032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,512,0.0031911111954185697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,8192,0.02172800070709652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,1024,0.004621333132187526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,6144,0.011379555695586734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,2560,0.03945688737763299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2048,512,0.014070222775141398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2560,65536,1.5495057635837133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,3584,0.009335111412737105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,6144,0.011003555523024665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,16384,0.029883556895785864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,10240,0.01204622205760744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,128,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,8192,0.01073155552148819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,6144,0.00997688869635264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,64,12288,0.0052133335007561584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,10240,0.148054215643141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,3584,0.01994311147265964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,7168,0.023686221904224817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,5120,0.015655999382336933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,16384,0.13871822092268202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,8192,0.19809422228071424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,2560,0.03696711195839776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,1536,0.014128888646761576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3584,65536,0.20925954977671304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,7168,0.05175111028883192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,3072,0.019729778170585632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,7168,0.03765155540572272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,1536,0.0069724445541699724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,7168,0.010010666317409938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,16384,0.16136711173587376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,6144,0.01750577820671929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,6144,0.03659111261367798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,7168,0.0122097780307134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,2560,0.042429334587521024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,6144,0.14860444598727757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,7168,0.05552266703711616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,3072,0.013409777647919126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,65536,0.06835466623306274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,768,0.003020444470975134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,6144,0.009000889129108852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,3072,0.32559111383226186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,2560,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,12288,0.013050666285885705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,128,0.0021768889079491296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,2560,0.014828445182906257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,3072,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,8192,0.010677333507272931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,16384,0.3654017713334825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,8192,0.020915556285116408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,512,0.0024871110088295406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,256,0.014128888646761576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,8192,0.018788443671332467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,16384,0.026176888081762526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,65536,128,0.02026844521363576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,6144,0.02150844368669722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,6144,0.21058044168684217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,10240,0.014399111270904541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,1024,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,2048,0.014731556177139282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,3072,0.007283555964628856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,512,0.006949333681000604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,5120,0.012433778080675336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,1024,0.003160888950030009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,12288,0.014440889159838358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,16384,0.26138756010267467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,6144,0.015806222955385845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,256,0.006513777706358168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,65536,0.06742133034600152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,10240,0.02537600033813053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,65536,768,0.025482666161325242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,64,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,1024,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,65536,1.276235580444336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,256,0.0530008872350057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,2560,0.004912000149488449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2048,4096,0.00797155582242542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,512,0.002620444529586368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,16384,0.044927110274632774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,5120,0.3258071210649279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,512,0.015423999892340766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,64,0.012051555845472546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,3584,0.022112889422310725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,768,5120,0.018208889497650992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,16384,0.027416000763575237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,7168,0.013755555782053204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,4096,0.02423111100991567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,3072,0.008239111138714684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,1536,0.007628444168302748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,16384,0.041590223709742226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,64,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,512,0.006288888967699475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,4096,0.0052586665583981415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,10240,0.03775999943415324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,512,7168,0.009782222410043081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,7168,0.3411288791232639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,3072,0.007678222325113084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,6144,0.018575999471876357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,8192,0.01346577786737018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,5120,0.020258666740523446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,512,0.021200888686709937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,2048,0.03350755572319031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,1536,0.008583111067612966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,512,0.00867555538813273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,10240,0.017879999346203275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,7168,0.035402665535608925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,8192,0.0047422221137417685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,8192,0.005193777796294955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,10240,0.11503822273678249
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,7168,0.014153778553009033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,512,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,5120,0.021321778496106465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,5120,0.01685066686736213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,16384,0.04858666658401489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,128,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,7168,0.011145778000354767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,7168,0.013655111193656921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,1536,0.011391999820868174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,512,0.05983822213278877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,65536,0.17880800035264757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,3072,0.007239110767841339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,65536,0.0732097758187188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,6144,2.749874750773112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,1024,0.07160444392098321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,768,0.005116444494989184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,8192,0.010695999695195092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,512,0.1160328918033176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,256,0.0028355556229750314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,1024,0.004905777672926585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,7168,0.011039999624093374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,16384,0.04119022356139289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,10240,0.012100444071822695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,128,0.009709333380063375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,12288,0.017196445001496207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,12288,0.10595377948549058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,512,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,8192,0.04206933246718513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,7168,0.012717333104875354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,128,0.005283555636803309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,8192,0.06253777609931098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,3584,0.01512888901763492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,3072,0.0151582227812873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,8192,8192,0.01885511146651374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,1536,0.006239999913507038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,5120,0.010001777774757808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,1536,0.010945777926180096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,512,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,128,0.01204622205760744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,65536,0.06401333543989393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,7168,0.031268444326188825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,64,0.002833777831660377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,64,0.0018568889548381169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,7168,0.013202666408485837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,32,0.023000889354281958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,5120,0.00834222220712238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,256,0.005221333354711533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,7168,0.03832088907559713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,1536,0.019559999306996662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,16384,0.020253333780500624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,512,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,256,0.005664888769388199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,65536,0.07956177658504911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,10240,0.2266871134440104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,6144,0.009327111144860586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,12288,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,64,0.0018773333479960759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,65536,16384,0.16916710800594756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,128,4096,0.005197333378924264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,128,0.006251555350091722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,256,0.00216977774269051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,7168,0.013976888524161445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,16384,3.0807147555881076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,8192,0.014479999740918478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,768,0.007527111305130853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,2048,0.023081777824295893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,10240,0.021884444687101576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,8192,0.01240533341964086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,16384,0.05796800057093302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,10240,0.051368888881471425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,32,0.0021297778520319196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,3584,0.7617546717325846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,3072,0.010471999645233154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,12288,0.00666222224632899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,32,0.009639110830095079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,10240,0.03216088811556498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,4096,0.03421599997414483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,1024,0.004877333425813251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,7168,0.009242666264375051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,3072,12288,0.03660355673895942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,10240,0.007269333634111617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,2560,0.006510222123728857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,7168,0.03152799937460158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,3072,0.04133066534996033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,128,0.03107644451989068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,65536,0.05824355284372965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,2560,0.008273777862389883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,128,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,2048,0.006609777609507243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,768,0.005604444278611078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,1536,0.024280000064108107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,4096,0.008056888977686564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,256,0.0021439999755885866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,6144,0.06332177586025663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,4096,0.044256889157825045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,768,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,5120,0.02392888896995121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,65536,0.36664266056484646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,1024,0.016174222032229107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,128,0.0052266667286554975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1536,2048,0.007649777664078607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,6144,0.1696995496749878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,2048,0.007082666787836287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,256,0.0025031111306614345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,5120,7168,0.012759110993809171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,512,0.01475644442770216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,2560,0.014446222119861178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,12288,0.02095111045572493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,7168,0.009648000200589498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,768,0.017838221457269456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,768,0.016148444679048326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,128,0.0052435555391841466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,3584,0.007599111232492659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,2048,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,2560,0.01572711103492313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,768,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,4096,0.009690666364298927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,2048,0.005888000130653381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,65536,0.35371822781032985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,1024,0.021295999487241108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,3072,0.01204711116022534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,10240,0.02324177821477254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,16384,0.005544888890451855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,8192,0.026337777574857075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,2560,0.01516266663869222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,5120,0.21903644667731392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,64,0.0018560000591807896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,512,0.008781332936551835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,4096,0.005828444328573015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,1536,0.016400888562202454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,3072,0.017864000466134813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,16384,0.04555733336342705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,4096,0.008344000412358178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,8192,0.07244710789786445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,4096,0.008672888908121321
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,65536,0.253256877263387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,3584,0.004861333303981357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,10240,0.013794666363133324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,1024,0.005617777920431561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,1536,0.008976000050703684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,65536,0.06827111376656426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,6144,0.027114666170544092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,256,0.008623999853928884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,12288,0.012699555191728802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,10240,0.03040177623430888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,12288,0.039408889081743025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,12288,0.034506665335761175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,65536,0.4197591145833333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,12288,0.09328977929221259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,1024,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,3584,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,1536,0.01551555593808492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,12288,0.010441777606805166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,512,0.007994666695594788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,2560,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,256,0.0052168890833854675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,7168,0.15944443808661565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,5120,0.004498666773239772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,32,0.010036444498433007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,1024,0.0074444446298811175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,512,0.01514933341079288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,12288,0.025394666526052687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,32,0.005229333208666907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,3072,0.03562666641341315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,32,0.028415110376146104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,3072,0.030841777722040813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,1536,0.012078222301271228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,512,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,2560,0.007679111427730984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,512,0.02161600026819441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,10240,0.01254666679435306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,5120,1536,0.008526222573386298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,3072,0.0070773329999711775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,32,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,16384,0.07126310798856947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,6144,0.01202933324707879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,2048,0.020572443803151447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,3072,0.02773511078622606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,3072,0.00742133292886946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,768,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,512,0.016930666234758165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,7168,0.02754755483733283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,4096,0.07042400042215984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,768,0.005560889012283749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,8192,0.024309333827760484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,3072,0.007415999968846639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,65536,0.13058488898807102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,768,0.0028080000645584534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,10240,0.021690666675567627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,16384,0.08512177732255723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,5120,0.016163556112183463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,65536,0.06486933098898993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,7168,0.01202933324707879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,6144,0.009335111412737105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,2048,0.009341333475377824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,7168,0.13064444065093994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,6144,0.01443199978934394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,5120,0.023518222901556227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,768,0.006065777606434292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,4096,0.039515554904937744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,12288,0.021363554729355708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,2048,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,128,0.018838221828142803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,128,0.007602666815121968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,8192,0.1968595584233602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,256,0.018236445056067575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,1536,0.02389777700106303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,128,0.0021057778762446507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,16384,0.04375822345415751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,64,0.005385777602593104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,3584,0.03379911184310913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,5120,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,3072,0.0840195549858941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,4096,0.023556444380018447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,8192,0.04314844475852119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,16384,0.021074665917290583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,6144,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,65536,0.050385776493284434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,65536,0.07214844226837158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,32,256,0.006583999842405319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,64,0.008223999705579545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,3072,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,32,0.0322746667597029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,12288,0.04086666636996799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1024,2560,0.004859555512666702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,3072,0.007617777420414819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,768,0.015657777587572735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,768,0.008255999949243333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,128,0.0052720002002186244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,7168,0.0233262214395735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,1536,0.008832889298597971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,3072,0.006928000185224745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,5120,0.01939733326435089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,65536,2048,0.07514578104019165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,5120,0.019348444210158456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,768,0.00702311098575592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,65536,0.23158044285244414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,512,0.014155555102560254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,4096,0.034407999780442976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,6144,0.011931555966536203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,768,0.00630044440428416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,65536,2.845343907674154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,5120,0.10686577690972222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,3584,0.0151591118839052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,8192,0.007239999870459239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,6144,0.016176889340082806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,3072,0.00888533310757743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,10240,0.16533154911465114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2560,5120,0.0225564440091451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,7168,0.014368888404634265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,6144,0.010018666585286459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,64,0.0021377777059872947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,1024,0.014833778142929077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,16384,0.030663109487957422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,10240,0.012118221984969245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,65536,0.13160622119903564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,768,0.0028764444092909494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,6144,0.050789333052105375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,6144,0.017882666654056974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,128,0.002945777856641345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,6144,0.009024000002278222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,768,0.007448888487286038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,768,0.008714666797055138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,6144,0.013344888885815939
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,1536,0.006246222390068903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,8192,0.014212444424629211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,64,0.0017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,16384,0.05917955769432915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,1024,0.007313778003056844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,128,0.002136888810329967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,5120,0.023714666565259297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,5120,0.013746666411558786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,7168,0.02608088817861345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,8192,0.03109955456521776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,12288,0.0433208876185947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,2560,0.03071911136309306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,1536,0.003530666646030214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,3584,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,512,0.021525333325068157
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,2560,0.009663110805882348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,512,0.00692622239391009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,8192,0.020581333173645865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,6144,0.014316444595654806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,5120,0.008970666262838576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,768,0.0028080000645584534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,5120,3072,0.02574311031235589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,1024,0.021426666114065383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,3584,0.011979555918110741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,4096,0.008303110798199972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,768,0.05986577934688992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,2560,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3584,256,0.0058871110280354815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,256,0.0021226666867733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,8192,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,5120,3072,0.00867911097076204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,1024,0.04200355543030632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,3072,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,1536,0.04850666721661886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,4096,0.012749333348539142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,6144,0.03185688787036472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,6144,0.023023999399609033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,512,0.012496888637542725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,2048,0.009670221971140968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,6144,16384,0.022740445203251306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,4096,8192,0.02527911133236355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,1024,0.015842666228612263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,256,0.07220444414350721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,2048,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,32,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,12288,2048,0.011342222491900126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,65536,0.21786933475070527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,256,0.029789331886503432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,768,0.006047111004590988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,2560,0.022484444909625586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,2560,0.007166222565703922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,16384,0.0440959996647305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,6144,0.02515022291077508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,16384,0.05597511265012953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,16384,0.03095910946528117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,7168,0.013758222262064615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,10240,0.01832266648610433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,512,5120,0.009706666900051964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,32,0.005583111196756363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,4096,0.04794666502210829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,512,0.016228444046444364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,3584,0.007680888805124495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,6144,0.02609777781698439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1024,1536,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,5120,0.010024000373151567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,2048,0.004587555511130227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,5120,0.017807111144065857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,12288,0.037858668300840594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,3072,0.026458667384253606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,4096,0.01753155555990007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,256,0.00216977774269051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,5120,0.028958221276601154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,5120,5120,0.03878577881389194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,768,0.01753422286775377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,4096,0.015136889285511441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,3584,0.007296889192528195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,3584,0.023297778434223596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,8192,0.011021333436171213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,256,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,2048,0.014733332726690503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,6144,0.1878133349948459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2560,10240,0.015817777978049386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,128,0.014479999740918478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,10240,0.0552462206946479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,65536,0.1749502287970649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,2560,0.07274311118655734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,1536,0.008366222182909647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,5120,0.01645066671901279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,2560,0.02915289004643758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,768,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,32,0.005537777725193236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,32,0.0021173333128293357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,16384,0.020940444535679285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,1024,0.009002666506502364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,256,0.012076444096035428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,2048,0.004221333397759331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,768,0.011680000358157687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,7168,0.028296000427669946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,2560,0.021983999345037673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,2560,0.0755866699748569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,4096,0.010575110713640848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,12288,1024,0.01034400032626258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,16384,0.04577599962552389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,8192,0.011021333436171213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,2560,0.08269066943062676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,128,0.014510222607188754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,8192,0.03899733225504557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,4096,0.010645333263609143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,3072,0.03159111075931125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,256,0.015440000428093804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,2560,0.021506667137145996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,256,768,0.005774222314357758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,12288,0.343470229042901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,6144,0.020222221811612446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,16384,0.04363822274737888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,256,0.017495110630989075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,16384,12288,0.04018755422698127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,768,0.01566133399804433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,4096,0.015812445018026564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,5120,0.008351999852392409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3584,1024,0.003519999898142285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,65536,7168,0.155967116355896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1536,7168,0.010811555716726514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,2560,0.01551644504070282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,1536,0.0038817777401871155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,1024,0.007469333708286285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,512,0.0059262220230367445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,2048,0.016166221764352586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,3072,0.007636444436179266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,10240,0.005577777822812398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3584,8192,0.01480977733929952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,32,0.002814222127199173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,65536,0.07681333356433444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,32,10240,0.012025777664449481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,6144,65536,0.20910044511159262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,8192,5120,0.43291023042466903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,5120,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,10240,0.01917155583699544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,256,0.002480888946188821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,6144,0.010834666589895884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,1024,0.009348444640636444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,64,0.002480000050531493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,65536,1.6068346235487196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,3072,0.007255111303594377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,7168,0.03246666656600104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,2048,0.01111288865407308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,10240,0.05942310889561971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,2048,0.018182223041852314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,65536,0.09353599945704143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,16384,0.04790488878885905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2048,6144,0.019912888606389362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,4096,0.020008888509538438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3072,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,1024,0.01573777695496877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,2560,0.02913244565327962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,4096,0.0433555543422699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,256,0.018222222725550335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,4096,10240,0.04217333263821072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,7168,0.023363555471102398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,7168,3584,0.027277333868874445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,256,0.005378666851255629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,768,0.0028248888750871024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,12288,0.03454400102297465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,128,0.006230222268237009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,12288,0.21975999408298066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,1024,0.00872622181971868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,5120,0.012028444144460889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,5120,0.01886666648917728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,7168,0.01973155637582143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,2560,0.020071110791630216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,128,0.005631999837027655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,8192,0.028607110182444256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,512,0.014112000664075216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,3072,0.007711110843552484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,4096,0.008284444610277811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,64,0.005672889037264719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,32,0.007583111524581909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,3584,0.006562666760550604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,6144,0.009540444446934594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,5120,0.00906577789121204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,10240,7168,0.044284443060557045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,5120,0.08451910813649495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,65536,0.08198044697443645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,12288,0.013064000341627332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,10240,0.09021333191129897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,128,0.005616888817813661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,6144,0.044010665681627065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,64,5120,0.008960888617568547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,1024,0.01348088930050532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,2048,0.009742221898502773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,1024,0.018237334158685472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,64,0.002052444343765577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,6144,0.03429333368937174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,32,1536,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,512,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,512,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,5120,0.08126133018069796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,768,0.0028355556229750314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,768,0.005963555640644497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,8192,0.011534222298198275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,7168,128,0.0047182221379545005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,4096,0.015999999311235216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,10240,0.014952888091405233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,128,0.0019511110666725372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,16384,0.03794577717781067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,5120,0.015972443752818637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,3584,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,2560,0.00490044429898262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,16384,0.4480995602077908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,3072,0.016384000579516094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,2048,0.011647111011876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,1024,0.007268444531493717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,65536,0.09549955526987712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,12288,0.020302222834693063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,2560,0.018262222409248352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,65536,32,0.01716800034046173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,1536,0.01772177716096242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,3584,0.007664889097213745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,3072,0.012648888760142855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1024,16384,0.02148444453875224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,3072,0.010388444695207808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,1536,0.008441777692900764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,512,0.015511110424995422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,128,0.005581333405441708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,16384,10240,0.05539466606246101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,2560,0.004569777597983678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,1024,0.018544000056054857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,16384,16384,0.07436622513665093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,256,0.0021662221600612006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,768,0.003485333174467087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,65536,0.0625822212960985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,256,0.012168000141779581
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,512,0.015153777268197803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,7168,0.01960444450378418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,6144,0.08005244202084012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,16384,0.13303466637929282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,12288,0.0206977774699529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,128,0.006413333531883027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,7168,0.052573333183924355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,1024,0.006589333216349284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,3072,0.01402666668097178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,1024,0.003863111138343811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2048,12288,0.04926222231653002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,5120,0.00488800017370118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,16384,0.013770666387346057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1536,3584,0.007291555404663086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,10240,0.026233777403831482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,5120,0.12455111079745823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,65536,0.10289244519339667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,512,0.005605333381228977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,8192,0.035994665490256414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,5120,0.008354666332403818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,1536,0.014466666513019137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1024,16384,0.019702222612169053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,8192,768,0.003530666646030214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,12288,0.024837333295080397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,12288,0.04263910982343885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,512,0.07939555247624715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,3072,0.012776888906955719
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,10240,256,0.012646222280131446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,32,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,512,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,7168,0.028065777487225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,12288,0.004861333303981357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,16384,0.011824000212881299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,6144,0.009335999687512716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,16384,0.11939467324150933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,12288,0.04060711132155524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,64,0.006650666809744305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,2560,0.04975822236802843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,2048,0.029115554359224107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,64,0.001800000046690305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,512,0.00998577806684706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,3072,0.008268444074524773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1536,16384,0.032884445455339216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,16384,0.01720266706413693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,3072,0.03073955575625102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,64,0.02271466619438595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,512,0.002187555655837059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,3072,0.014515555567211576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,32,0.006296000132958095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,65536,0.16177422470516628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,10240,0.018258665998776753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,2048,0.004224888980388641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,3584,0.01963911122745938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,32,0.0058853332367208265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,768,0.02994489007525974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,3072,0.029162665208180744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,16384,0.07979644669426812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,6144,0.009752000371615091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,64,0.0018497777895794974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,128,0.015037332971890768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1536,3584,0.007643555601437886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,10240,0.04349955585267809
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,1024,0.010775110787815519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,64,5120,0.00867555538813273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,5120,0.020615110794703167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,65536,6144,0.05574311150444878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,256,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,32,12288,0.025087111526065405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,512,0.005607999861240387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,3584,0.0199644449684355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,128,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,8192,0.015504888362354703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,12288,0.5842284626430935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,2560,0.022658665974934895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,10240,0.018124444617165458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,65536,0.13426221741570368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,1024,0.021984888447655573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,768,0.024001777172088623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,12288,0.06839822398291694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,65536,0.20786844359503853
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,10240,0.025519111090236243
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,5120,0.10917511251237656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,7168,0.01605333387851715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,16384,5120,0.045145776536729604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,128,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,2560,0.0076044441925154785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,768,0.0024862223201327855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,1536,0.006278222219811545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,6144,0.019879110985332064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,128,0.016238222519556682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,10240,0.01590666671593984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,10240,0.031537777847713895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,3584,0.011173333558771344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,8192,0.03956977857483758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,6144,0.023611555496851604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,256,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,3072,0.0192586663696501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3072,3072,0.009083555804358589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,256,0.01551555593808492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,32,0.0018622221218215094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,12288,10240,0.8928417629665799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,32,8192,0.011697777443461947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,65536,0.22880710495842826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,32,6144,0.009708444277445475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,256,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,2560,0.018014222383499146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,12288,0.039020445611741804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,8192,0.13741066720750597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,1536,0.006608888920810487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,1536,0.03242133392228021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,768,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,12288,0.015729778342776828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,65536,32,0.029247111744350855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,128,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,1536,0.023165333602163527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,4096,0.00833511104186376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,6144,0.009390222529570261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,8192,0.04255911045604282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,65536,0.4406248993343777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,10240,0.09898933437135483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,1024,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,1536,0.0165928883685006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,3584,0.014084445105658637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,12288,0.06920444303088717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,128,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,512,0.002941333378354708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,2560,0.009101333717505137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,65536,12288,0.24623288048638237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,1536,0.008656000097592672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,3584,0.010336000058386061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,64,0.005582222094138463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,32,0.010663111176755695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,12288,0.02957688768704732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,3072,0.004528000122971005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,768,512,0.015152000718646579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,1024,0.007272000114123027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,2048,10240,0.18663733535342744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,256,0.005234666582610872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,65536,0.1400968896018134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,768,0.006254222244024277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,5120,128,0.005932444499598608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,65536,0.05645333396063911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,10240,0.10997866921954685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,2560,0.0124177775449223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,3584,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,12288,0.02784622377819485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,10240,0.005565333283609814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,65536,0.23748177952236602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,1024,0.010379555324713389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,2560,0.008771555291281806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,10240,0.02879200047916836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,10240,0.12117866675059001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,1536,0.0035297779573334586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,16384,0.007296889192528195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,16384,1.506765365600586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,5120,0.011753777662913004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,3072,0.01312711089849472
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,3072,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,7168,0.0068959999415609576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,128,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,3584,0.022689777943823073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,7168,0.013339555925793119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,5120,0.004921777794758479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,256,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,32,0.0018382221460342407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,16384,128,0.002130666747689247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,10240,0.01401422255569034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,512,0.006615110983451207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,3072,0.009305777649084726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,5120,0.016599999533759225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,2048,0.01790488925245073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,2560,0.03689333465364244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2560,65536,0.06562399864196777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,512,7168,0.009743111001120673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,5120,7168,0.016418667303191293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,2560,32,0.0018595556418100994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,10240,0.01719466679626041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,8192,0.01925244430700938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,8192,0.011706666813956367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,65536,1.5105519824557836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,3584,0.27232710520426434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,256,0.00488711107108328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,5120,0.015651555524932016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,3072,0.015609777635998197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,2560,0.015826667348543804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,3072,0.07417688767115276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,2048,0.015840889679061044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,65536,0.025063999825053748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,3584,0.025069332785076563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,4096,0.02931555443339878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,1024,256,0.008661333057615492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,10240,65536,0.1395688851674398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,3072,0.07268444697062175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,64,8192,0.07012355327606201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,512,0.002809777855873108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2048,8192,0.023757333556811016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,16384,0.024978665841950312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,2048,0.025638222694396973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,768,2048,0.006596444381607904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,10240,0.025703110628657874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,1536,0.007239110767841339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,8192,0.038367112477620445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,10240,0.03840177920129564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,3072,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,3584,0.030393779277801514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,5120,0.009378666679064432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,8192,0.015400888191329109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,2560,0.024338665935728285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,3584,0.006934222247865465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,1024,0.00627377794848548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,768,0.0025537777692079544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,8192,0.03390222125583225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,256,768,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,512,0.002528888897763358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,2048,0.012053333222866058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,128,0.005200888961553574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,2560,0.011752000285519494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,7168,0.012799111505349478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,512,0.005221333354711533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,128,0.03678755627738105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,8192,0.025472889343897503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,3072,0.010699555277824402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,256,0.02507200009293026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,128,0.005062222066852781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,4096,0.011696000066068439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,768,5120,0.009329777624871995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,6144,0.13046844800313315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,16384,0.04941688974698385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,6144,0.023077333966890972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,10240,0.015772443678643968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2560,2560,0.006956444846259222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,512,0.0028640000770489373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,16384,0.1400702264573839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,768,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,6144,0.013607111242082385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,4096,0.007982222570313348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,128,0.0042035554846127825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1536,65536,1.0717013676961262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,8192,0.09168622228834365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,7168,0.018629334039158292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,3584,0.0073831114504072405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,16384,0.044032888280020825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,5120,0.027349332968393963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,256,0.01593866613176134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,6144,8192,0.016895999511082966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,2048,0.012072888513406118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,7168,1536,0.02169688873820835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,6144,8192,0.014757333530320061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,7168,8192,0.023581332630581323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,12288,0.01350044459104538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,2048,0.006994666324721442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,1536,0.006492444624503453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1536,128,0.01479200025399526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,1536,0.0038622220357259116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2048,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,3072,0.010038221875826517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,512,0.005606222069925732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,64,32,0.0021333334346612296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,16384,0.015776889191733468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,256,0.021189333664046392
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,3584,0.014143110977278816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,1536,0.006242666807439592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3072,7168,0.030101332399580214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,512,4096,0.006649777707126405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,12288,0.013746666411558786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,512,0.015288889408111572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,12288,0.009898666706350114
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,4096,0.01906844476858775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,768,0.016809778081046212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,3584,0.010994666980372535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,64,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,2560,0.010034666293197209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,256,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,256,0.005250666704442766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,16384,0.9000444412231445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,128,0.05680533250172933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,3072,0.013472889032628802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,2560,0.00944800012641483
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,256,0.005597333527273602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,65536,0.07804977893829346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,768,0.08363822433683608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,32,1024,0.002875555513633622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,16384,0.021184000704023574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,768,0.003212444484233856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,16384,0.015783111254374187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,10240,0.02481511069668664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,7168,0.03796266516049703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,3584,0.04002666804525588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,512,0.03756533397568597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,128,0.002472888885272874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,512,0.007085333267847697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,3584,0.08513155248430039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,65536,0.14368444018893772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,2048,0.010312889185216691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,128,0.022678222921159532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3584,1024,0.01585244470172458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,65536,0.4794044494628906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,10240,0.016697777642144095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,128,5120,0.009026666482289633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,512,0.006319111006127463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,5120,0.012450666891203986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,64,0.005965333431959152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,1536,0.20024711555904814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,12288,0.01072355525361167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,768,0.01480977733929952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,128,0.00574222207069397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,12288,0.025045333637131587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,3584,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,256,0.017831999394628737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,256,0.01647733317481147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,1024,0.0069271110826068455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,128,0.022896000080638464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,1536,0.013777777552604675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,128,0.0063182223174307085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,12288,0.05013244350751241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,6144,0.016173332929611206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,12288,0.32571289274427623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,2048,0.016922665966881644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,4096,0.0158924443854226
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,5120,0.04660177893108792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,65536,3584,0.03485155436727736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,128,0.0028195555011431375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,5120,0.021066667305098638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,3584,0.03039111031426324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,5120,0.024762666887707178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,7168,1536,0.006248888870080312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,6144,0.03459555572933621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,768,0.048327111535602145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,64,0.003863999827040566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,16384,0.017080000705189176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,32,0.012055111428101858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,32,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,2048,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,16384,0.37269067764282227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,2560,0.005203555441564984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,1024,0.0031902222997612427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,3584,0.031360887818866305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,6144,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,2560,0.06062666575113932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,10240,0.017447999782032438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3072,5120,0.008626666333940294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,2560,0.009313777916961247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,1024,0.07761777771843804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,4096,0.008195555872387355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,512,0.014817777607176038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,7168,0.012849777936935425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,10240,0.012196444802814059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,4096,0.17029421859317354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,3584,0.04156977931658427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,64,0.012200000385443369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,2048,0.006991110742092133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,64,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,7168,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,3072,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2560,16384,0.021697777840826247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,7168,2560,0.007715555528799693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,2560,0.013399110900031196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,1536,0.00813600007030699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,1536,0.011401777466138205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,32,0.0018515555808941524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,3584,0.017846221725145977
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,768,0.014337778091430664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,2560,0.007596444752481248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,8192,0.02362577782736884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,1536,0.0443484452035692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,10240,7168,0.06121244695451525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,2560,0.017895110779338412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,8192,0.01872711049185859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,7168,0.013640888863139682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2560,256,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,65536,0.047000888321134776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,64,0.0018168888572189545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,8192,0.039507554637061224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,256,0.005623111294375525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,128,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,1024,0.007722666694058313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,65536,0.1328266726599799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3072,16384,0.021456888980335657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,5120,0.02681244413057963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,12288,0.019536889261669584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,12288,0.06388266881306966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,8192,0.033324443631702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,768,0.003519111209445529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,1024,0.011626666618718041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,65536,0.07118133041593763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,64,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,768,32,0.006631111105283101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,16384,2048,0.012753778033786349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,256,1024,0.012442666623327466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,6144,10240,0.019534221953815885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,16384,0.03435644507408142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,3072,0.03223022156291538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,512,0.009338666995366415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,4096,0.017822222577200997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,2560,0.020211555891566806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,2048,0.007600000335110559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,32,0.0023164444913466773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,512,0.006634666687912411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,8192,0.008551999926567078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,512,0.011832888755533429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,64,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,65536,0.07454488674799602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,768,0.018633777896563213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,256,0.00656444455186526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2560,768,0.007277333074145847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,6144,0.011083555718262991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,5120,0.007282666862010956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,8192,0.032771554258134626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,4096,0.011398221883508893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,2560,0.006888888776302338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,10240,0.012395555774370829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,128,0.0358933342827691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,4096,0.016520000166363187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,128,0.0069217777086628815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,65536,10240,0.10129600101047093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,64,0.002179555594921112
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,16384,0.01733955575360192
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,512,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,128,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2048,65536,0.06390755706363253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,4096,16384,0.022263111339675054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,12288,0.012418666647540199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,256,0.024471999870406255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,1536,0.00813333359029558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,10240,0.0844106674194336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,8192,0.04660977919896444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,65536,0.12319289313422309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,32,0.0018195555441909367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,1024,0.017789334058761597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,3072,0.007680888805124495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,128,16384,0.012183999849690331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,3584,0.007623111208279927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,512,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,256,0.005620444400442972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,128,0.002477333363559511
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,10240,0.4812240070766873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,5120,65536,0.08015910784403484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,512,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,7168,0.022802667485343084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,768,0.005386666705211003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,4096,0.043656001488367714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,6144,0.01612622208065457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,6144,0.04655111167165968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,65536,0.06733510891596477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3072,7168,0.01111822244193819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,7168,0.011047110789351992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,64,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,12288,0.01406133340464698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,6144,0.011904888682895236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,1536,0.015486222174432544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,16384,0.0247688889503479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,4096,0.01905422243807051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,1536,0.06694044669469197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,1024,0.01517688896920946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,16384,0.020690666304694284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,5120,0.008651555412345463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3584,256,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,128,0.005535111245181825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,8192,0.04279911186960009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,3584,0.014833778142929077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,7168,0.010031999813185798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,768,0.01480888823668162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,65536,0.09234755569034153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,5120,0.05893955628077189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,768,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,768,5120,0.010125333236323463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,8192,0.011721777419249216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,256,0.009395555489593083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,4096,0.226548433303833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,7168,0.021760000122918025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,2560,0.007599111232492659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,128,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,16384,10240,0.23758488231235078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,7168,0.029306666718588933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,256,0.0022826666633288064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,2048,0.01681866745154063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,10240,0.011199110911952125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,10240,256,0.002485333424475458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,65536,0.05678666962517632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,3072,0.01034577770365609
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,12288,0.08611466487248738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,1024,0.006602666858169768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,65536,0.064882669183943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,16384,0.056886222627427846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,16384,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,768,0.009353777600659264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,3584,0.015087111128701104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,4096,0.007865777446164025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,5120,0.0587751136885749
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,16384,128,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,16384,0.023464888334274292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,5120,0.016473778420024447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,65536,3.2842896779378257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,12288,0.07672444317075941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,10240,0.030497776137457952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,4096,0.008963555925422246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1536,2560,0.015787555111779105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,512,0.0041928887367248535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,256,0.002528888897763358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,64,0.005236444373925527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,2560,0.009988444546858469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,2048,0.0069057775868309875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,5120,0.011004444625642566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,12288,0.11412089400821263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,65536,0.06755822234683566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,512,0.019389332996474374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,64,0.0022177776942650476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,2560,0.007263110743628607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,1536,0.0052151112920708126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,65536,0.0214684440029992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,32,6144,0.004925333377387789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,4096,0.018073777357737224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,10240,0.012048888537618848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,6144,0.011605333122942181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,64,0.006846222198671765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,3072,0.010033778018421596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,7168,2048,0.028773334291246202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,4096,0.01400888959566752
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,512,3584,0.007635555333561367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,128,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,512,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,10240,10240,0.5655395719740126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,7168,0.021949332621362474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,16384,12288,0.043825778696272105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,65536,0.15041155285305446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,2048,0.00865955568022198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,512,0.0047306666771570844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,10240,0.029709332519107397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,5120,0.018552000323931377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,65536,0.06704977485868666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,32,128,0.008151999778217739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,4096,0.010183999935785929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,12288,0.02128177715672387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,6144,0.016490666402710807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,6144,0.018216000662909616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,128,0.004920888692140579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,8192,128,0.015458666616015963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,6144,0.021007110675175984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,512,0.028022223048739966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,5120,0.14022221830156115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,2048,0.05433511071734958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,2048,0.009316444396972656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,2560,0.004902222090297275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,8192,8192,0.047080000241597496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,10240,0.011809777882364062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,128,0.0021128888345426982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,768,0.003134222287270758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,3584,0.007424888511498769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,65536,0.15277243985070124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,2048,0.018180444836616516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,7168,0.12358310487535264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,8192,0.06996888584560819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,65536,0.013415111435784234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,128,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,7168,0.0311173333062066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,64,0.0027591110103660156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,512,0.009198221895429824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,1024,0.00350577798154619
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,4096,0.033666667011049055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,12288,0.01982666717635261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,5120,0.00979911122057173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,4096,0.017668444249365065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,128,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,768,0.0028684443483750024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,16384,0.05565600262747871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,2560,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,128,0.002089777754412757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,2560,0.019471999671724107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,1536,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,5120,0.15785600079430476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,768,0.005933333188295364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,65536,0.15148622459835476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,32,0.006218666831652324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2560,7168,0.022658665974934895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,3584,0.5916293462117513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,10240,0.012362666428089142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,1536,0.003727999826272329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,8192,0.08870133426454331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,12288,0.198600000805325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,5120,0.028154667880800035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,2048,0.016826666063732572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,65536,0.14471733570098877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,4096,0.00885511106914944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,3072,0.011891555454995898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,3584,0.014319999350441827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,7168,0.004947555561860402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,12288,0.015505777464972602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,32,6144,0.009350222018029954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,128,0.006219555520349079
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,128,0.002463999928699599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1024,1536,0.0038399998512532976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,3584,0.023914666639433965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,1024,0.014163555370436775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,8192,0.010681777364677854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,768,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,1024,0.00318755561278926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,128,0.0021235555824306277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,4096,0.009081777599122789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,65536,0.06240089072121514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,3072,0.028410666518741187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,16384,0.006238222122192383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,512,0.004182222402758068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,3072,0.009959110783206092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,768,0.017197334104114108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1536,768,0.004569777597983678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,6144,0.043014221721225314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,3072,0.010850666297806634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,2048,0.007304000357786815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1536,4096,0.017831999394628737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,4096,0.008952000074916417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,65536,0.5273573133680556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,6144,0.1382008923424615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,6144,0.017429333594110277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,1536,0.026826666461096868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,1024,0.04627644353442722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,1024,0.02611555490228865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,6144,128,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,10240,0.3456497722201877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,5120,0.031311111317740545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3584,64,0.014136888914638095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,128,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2560,8192,0.011014222270912595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,65536,7168,0.10753422313266331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,12288,0.21791199843088785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,3584,0.005877333382765452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,7168,0.030448890394634668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,8192,0.031186666753556993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,64,0.02640977833006117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,32,0.005192000004980299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,7168,0.7818231052822537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,2560,0.007963555554548899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,7168,0.028985778490702312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,2048,0.004210666649871402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,2560,0.007131555842028723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,128,0.005924444645643234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3072,7168,0.049401776658164136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,65536,0.06602578030692206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,10240,0.04487555556827121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,3072,8192,0.029655999607510034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,7168,0.017246221502621967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,7168,768,0.010461333725187512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,16384,0.036590221855375506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,4096,0.008775110873911116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,3072,0.01647111111217075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,7168,0.009840000006887648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,8192,0.014560000763999091
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,7168,0.025472000241279602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,7168,0.17253422737121582
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,8192,0.015320888823933072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,16384,0.02198311189810435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,8192,10240,0.05794044335683187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,1536,0.00630666646692488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,1536,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,3584,0.038704888688193426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,10240,0.0162444445821974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,65536,0.2677875624762641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,128,0.005246222019195557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,512,0.029049777322345312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,1024,0.019561777512232464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,1536,0.007072000039948358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,65536,0.0653066635131836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1024,2048,0.007280000381999546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,4096,0.06103999747170342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3584,256,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,256,0.02260533306333754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,32,0.004872888740566042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,5120,0.017041777571042378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,16384,0.02307644486427307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,12288,0.01837688850031959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,768,768,0.016181333197487723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,6144,0.08124711116154988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,8192,0.016568889220555622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,6144,0.0123848890264829
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,65536,0.7862133449978299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,5120,0.008344888687133789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,65536,0.31994933552212185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,64,0.00519822206762102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,2560,0.025423111187087163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,2560,0.02921688887808058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,1536,0.006295111030340195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,7168,0.05446666810247633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,1536,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,4096,0.048130667871899076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,12288,0.027087110612127516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3584,10240,0.015190222197108798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,8192,0.005445333404673471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,16384,0.015547555353906421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,3072,0.007325333025720384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,16384,0.0258542216486401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,7168,0.022848000129063923
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,10240,0.04355555441644457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,128,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,65536,0.1280168824725681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,128,0.014119999276267158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,65536,8192,0.1743253337012397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,5120,2560,0.008014221986134848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,16384,0.02205955485502879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,6144,0.016163556112183463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,65536,0.16296088695526123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2048,65536,0.0639315578672621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,128,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,5120,0.012372444073359171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,10240,0.028127110666698877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,65536,128,0.012075555821259817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2560,512,0.013974222871992322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,2560,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,2048,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,7168,0.00964355551534229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,8192,3072,0.010327111515733931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,768,0.013784888717863293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,12288,0.05610399776034885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,5120,4096,0.019111999207072787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,768,5120,0.009020444419648912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,8192,0.015318221516079374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,7168,0.013783999615245394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,10240,0.020239111449983384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,256,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,16384,0.08176800277498034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,12288,0.04145333170890808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3584,3584,0.010310221877362993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,7168,1024,0.008358221915033128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,1536,0.006738666858938005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,1536,0.00868444475862715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,2560,0.0069573331210348345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,3584,0.04155200057559543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,512,512,0.006230222268237009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,10240,512,0.018534221582942538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,6144,0.01701955497264862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,3584,0.051303111844592623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,10240,0.037449777126312256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,1024,0.07441066371070014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,6144,0.03507733345031738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,1024,0.013456888496875763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,2048,0.0068986668354935115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,1024,0.012406222522258759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,1024,0.007589333587222629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,12288,0.013412444127930535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,768,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,65536,0.18937244680192736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,768,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,3072,0.006218666831652324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,256,0.004522666749027041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,10240,0.024033778243594702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,16384,1536,0.03405066662364536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,10240,5120,0.020871111088328894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,2560,0.007701333198282454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,1536,0.03350399931271871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,2048,0.009361777868535783
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,768,1024,0.00794222205877304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,64,8192,0.04040266738997565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,10240,0.013453333742088742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,8192,0.0240328891409768
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,768,2560,0.020773332979944017
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,5120,0.009004444711738164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,2560,0.0075866662793689305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,8192,0.012423111332787408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,768,0.0053084443012873335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,512,0.048156443569395274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,5120,3584,0.009703111317422655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,1536,0.009944889280531142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,1024,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,12288,0.0306915541489919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,128,0.00795377790927887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,6144,0.01793955597612593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,8192,0.014136888914638095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,4096,0.02859555681546529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,7168,0.011400000088744693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,6144,12288,0.019863110449579027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,2048,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,12288,0.023927110764715407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,16384,768,0.006000888844331105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,512,0.054971555868784584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,6144,0.010690666735172272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,5120,0.008623999853928884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,4096,0.020469332734743755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,3584,0.01442399952146742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,65536,0.10305866930219863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,64,32,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,4096,12288,0.018561777141359117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,3072,0.02607466611597273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,3584,0.00869599978129069
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,10240,128,0.006010666489601135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,16384,0.011335999601417117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2560,6144,0.009078222016493479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,768,0.01754844519827101
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,2560,0.009019555317031013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,10240,0.06234755780961779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,65536,0.16994755797915986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,3584,0.012044444680213928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,5120,0.03247022297647264
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,512,0.013079110946920184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,2560,0.013058666553762225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,1024,0.00794577764140235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,10240,0.036115556955337524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,768,0.01923022170861562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,7168,0.01594044433699714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,3072,0.025076443950335186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,512,0.015915556086434256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,65536,0.901942253112793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,1536,0.007744889292452071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,8192,7168,0.013724444641007317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,128,0.005557333429654439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,5120,0.03864088985655043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,6144,0.009290667043791877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,2048,0.00424888895617591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,1024,0.0028124445428450904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,6144,10240,0.018851555056042142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,6144,0.008984888593355814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,1536,0.008359111017651027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,5120,0.009660444325870937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,64,0.002858666703104973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,512,0.04136000076929728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,1536,0.02683999968899621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,2560,0.029410666889614526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,128,0.012103999654452005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,1024,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,2560,0.012082666986518435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,16384,0.06631377670500013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,12288,0.03297777639495002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,4096,0.02300888962215847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,6144,0.010002666877375709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,3584,0.007314666277832455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,768,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,8192,0.013737777868906656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,12288,0.038980444272359215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,6144,0.06910577747556898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,8192,0.01069866700304879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,2048,0.010701333483060202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,64,3584,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,10240,0.01588000026014116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,768,8192,0.0088755554623074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,5120,8192,0.015817777978049386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,1536,0.011061333119869232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1024,128,0.0021679999513758556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,6144,0.010058666268984476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,10240,1024,0.020949333906173706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,2048,0.021597334080272253
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,64,0.0021119999388853707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,7168,0.19890933566623262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,2048,0.016525333126386006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,1536,0.004193777839342753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,128,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,3072,0.022169777088695105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,2048,0.004536889079544279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,256,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,32,2048,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,768,0.017775111728244357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,1024,0.013087111214796701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,3584,0.009321778184837764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,6144,1024,0.0052328887912962176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,4096,0.026521777113278706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,16384,0.04434222314092848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,3584,0.011721777419249216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,1024,0.011713777979214987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,10240,1.496582243177626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,10240,65536,0.13978044191996256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,768,0.024713777833514746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,12288,0.013432888521088494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,8192,0.3200933403438992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1536,256,0.008551999926567078
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,3072,0.007629333270920648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,1024,0.014449778530332776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,6144,0.01791466606987847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,256,0.019463999403847587
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,4096,0.008045333127180735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,2560,0.14325866434309217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,5120,0.03744533326890733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,3584,0.014226666755146451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,2560,0.004914666629499859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,512,0.0034844444857703317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,4096,0.006066666709052191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,1024,0.00315022220214208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,16384,0.04667288727230496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,4096,0.08932089143329197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,3072,0.013307555682129331
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,8192,0.013839999834696451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,6144,0.011952888634469775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,7168,0.025941333836979334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,768,0.01651733285850949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,8192,0.03200088938077291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,7168,0.02408711115519206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,8192,0.03662933243645562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,6144,0.05965422259436714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,65536,1536,0.06807733244366117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,64,5120,0.004573333180612988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,512,0.00926755534278022
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,16384,0.21197866068945992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,7168,0.01107377807299296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,1536,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,1536,0.009730666875839233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,1024,0.006311111152172089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,512,0.014436443646748861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,3072,0.00423555572827657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,7168,16384,0.03906933466593424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,128,0.0041804446114434134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,4096,0.048037333620919116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2048,2560,0.018358222312397428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,16384,0.00554666668176651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,64,65536,0.10825600226720174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,32,0.007898666792445714
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,2048,0.005048888838953442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,512,1536,0.0059368887709246735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,4096,0.03123377760251363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,768,7168,0.025695110360781353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,1536,0.030065778228971694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,6144,64,0.0018026667336622875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,2560,0.005179555465777715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,2048,0.01585155559910668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,10240,0.01204177737236023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,8192,0.024471999870406255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3584,64,0.011572444604502784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,128,0.005218666460778978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,128,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,3072,0.01440533333354526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,2048,0.006621333460013072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,512,0.0031866667171319327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,512,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3584,2048,0.0069208890199661255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,128,0.003944888710975647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,2560,0.007241778075695038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,3072,0.01293688929743237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,6144,0.011356444822417365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,512,0.0028159999185138275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,12288,3584,0.015964445140626695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,768,0.005940444353553984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,64,0.005578666511509154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,4096,768,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,4096,0.011722666521867117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,128,0.001847111102607515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,64,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,1536,0.006551110910044775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,6144,0.04654755526118808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,6144,0.016739555531077914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,65536,0.7722213533189562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,128,512,0.017080888152122498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,5120,0.008018666671382057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,128,0.007916444705592262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,6144,0.026017778449588354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,65536,0.10725955830679999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,2560,0.00997511131895913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,512,0.12290755907694499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,4096,0.01070844464831882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,10240,0.015274667077594332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,5120,0.03594400154219733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,7168,0.018204443984561496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,256,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,3072,0.012536889149083031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,10240,0.03420089019669427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,4096,0.007939555578761632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,4096,0.005191111316283544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,512,0.018199111024538677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1536,1536,0.017098666893111337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,6144,0.014074666632546319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,32,0.0034888887570963967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,2560,0.00923555592695872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,7168,0.010328888893127441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,32,0.005899555567238066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,7168,0.20446577337053087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,5120,0.009371555513805812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,10240,0.01630311045381758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,32,0.005566222386227713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2048,3072,0.016150222884284127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,7168,0.010039110978444418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,12288,0.018373333745532565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,2560,0.07498221927218966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,32,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,768,0.005748444547255834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,6144,0.027999109692043726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,2560,0.01313244468635983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,8192,0.013186666700575085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,256,16384,0.016177778442700703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,3072,0.016488000750541687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,6144,0.008085333638721043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,3072,0.007326222128338284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,12288,0.018618666463428073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,2560,0.03190133306715224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,256,0.006406222366624408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,12288,0.04088000125355191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,2560,0.008012444608741337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,2560,0.007626666790909237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,16384,0.16683022181193033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,10240,0.016233777006467182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,512,0.042797333664364286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,10240,0.014540443817774454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,3072,0.009333333207501305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,2560,0.005264889034960005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,16384,4096,0.01575555569595761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2048,32,0.0083155557513237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,3072,0.00555288874440723
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,256,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,16384,0.3297688961029053
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,512,0.012814222110642327
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,65536,2560,0.15349422560797796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,6144,0.009362666971153682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,256,0.005229333208666907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,10240,0.016252444850073922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,1536,0.0059075554211934405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,768,0.005910222315125995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,2048,0.014466666513019137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,512,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,12288,0.005193777796294955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,10240,0.012040889097584618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,3072,0.00702311098575592
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,256,0.007802666889296637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,7168,0.03199911117553711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,8192,2560,0.00868000007337994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,32,0.005561777700980504
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,1024,0.018010665973027546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,2560,0.007292444507280986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,512,0.015468445089128284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,5120,0.008278222547637092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,4096,0.008414222134484185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,65536,0.027422222826215956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,256,768,0.005602666487296422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,256,0.0028391112056043413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,7168,0.017229333519935608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,256,12288,0.013316444224781461
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,10240,0.016600888636377122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,4096,0.029469334416919287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,6144,16384,0.06335110796822442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,7168,0.005247999810510212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,1536,0.004400889078776042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,512,65536,0.0625786648856269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,1024,0.007306666837798224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,3072,128,0.0059279998143513995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,256,0.028402666250864666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,10240,0.012730666332774691
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,2560,0.008329778081840938
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,6144,0.05407111181153191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,2048,0.0052577778697013855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,3072,0.009286222358544668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,16384,0.047261334127850003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,16384,0.01648533344268799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,512,0.005929777605666055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,64,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,6144,0.010004444254769219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,512,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,128,0.01428533262676663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,2048,0.016384000579516094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,10240,0.020937777227825586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,12288,256,0.0063688887490166565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,2560,0.006945778098371293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,512,0.005724444571468566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,2048,0.031148443619410198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,3072,0.007947555846638149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,10240,0.01442399952146742
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,65536,0.28945332103305393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,5120,0.013772444592581855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,768,0.003528000166018804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,8192,0.025306666890780132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,7168,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,5120,0.010527110762066312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,12288,512,0.008616888688670265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,8192,0.021410667233996924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,16384,0.21018044153849283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,768,0.0029395555870400537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,5120,0.02996088729964362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1024,16384,0.019110222657521565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,6144,0.01068800025516086
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,65536,0.04903644323348999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,32,0.010627555350462595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,8192,0.023399111297395494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,16384,0.3302995628780789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,256,0.005951111101441913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,3072,0.022852443986468848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,10240,0.040008889304267034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,3584,0.027391110857327778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,2560,0.008273777862389883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1536,3584,0.007617777420414819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,7168,0.025058666865030926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,7168,0.013447110851605734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,16384,2560,0.0326222214433882
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,768,0.0038488888078265717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,512,0.014675555957688225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,128,0.005542222410440445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,128,0.014084445105658637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,2048,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,7168,0.013321778012646569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,8192,2048,0.00795555528667238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,768,0.02126577827665541
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,4096,1024,0.014477334088749357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3072,6144,0.012996444271670448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,16384,3072,0.01647555496957567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1024,2048,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,7168,0.02124622298611535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,128,16384,0.016519111063745286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,3072,0.011711111499203576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1536,12288,0.014309333430396186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,7168,0.010513777534166971
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,10240,0.17851022879282633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,12288,0.01685599982738495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,7168,0.022613333331214056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,1536,0.01715022159947289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,32,3072,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,10240,0.015012444721327888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,2048,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,2560,0.007637333538797166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,12288,0.03157777918709649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,128,2560,0.023289778166347083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,3072,0.00795022232664956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,8192,0.026597332623269822
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,128,0.005548444473081165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,6144,0.0163137780295478
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,5120,0.017814222309324477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,128,65536,0.05351288782225715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,4096,5120,0.01701244380739
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,3072,0.03647111190689935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,128,0.015457777513398064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,2560,0.014415999253590902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,128,0.005576000031497743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,16384,0.2242844369676378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,16384,0.040580444865756564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,1024,0.0028231110837724474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,65536,0.04690222276581658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,512,0.013985777894655863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,768,3072,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,4096,65536,0.07092444764243232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,7168,0.01682577696111467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,16384,3584,0.03869422276814779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,2048,0.021384000778198242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,768,0.017441777719391715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,768,7168,0.012401777837011548
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,768,0.18600355254279244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,128,0.0021946666141351065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,128,32,0.005600000007285013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,16384,1024,0.02681155502796173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,1536,0.020210666788948905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,6144,0.018062222335073683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,1536,0.01442577772670322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,16384,0.025042666329277888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,64,0.0055440002017551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,128,0.007254222200976477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,65536,3.6376931932237415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,768,0.006232888748248418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,4096,0.027664888236257765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,65536,12288,0.25742578506469727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,3072,0.05013599991798401
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,768,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3072,256,0.006047111004590988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,12288,0.07936000161700778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,5120,0.18182044559054902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,6144,5120,0.014127110441525778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,6144,0.06844000021616618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,512,0.01239377756913503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,4096,0.006263111200597551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,1536,0.0038417776425679526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,3072,0.016542222764756944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,10240,0.013409777647919126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1536,7168,0.03046222196684943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,3584,0.040644443697399564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,3584,0.007613333563009898
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,2560,0.009399111072222391
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,6144,0.011020444333553314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,2048,12288,0.028382221857706707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,65536,0.07627644141515096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,256,5120,0.009309333231714036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,7168,0.014152889450391134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,128,0.0018666666001081467
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2560,7168,0.02663199934694502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,5120,0.011874666644467248
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,4096,12288,0.020236444142129686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,768,0.23616266250610352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,6144,0.013743110828929476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,12288,0.020741333564122517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,8192,0.05118044548564487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,12288,0.018762666318151686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,12288,0.027776888675159876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3072,768,0.01684799955950843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,1024,0.012069332930776807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2048,5120,0.027123555541038513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,32,0.0025235555238193935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,3072,0.03516799873775906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,7168,0.016113777955373127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,2048,0.026728888352711994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,1536,0.019528888993793063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,3072,0.023937778340445623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,2048,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,65536,768,0.017142222987280954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,3584,0.007311111523045435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,16384,2560,0.017882666654056974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,128,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,16384,0.03427377674314711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,256,0.007364444434642792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,7168,0.009693333672152625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,768,0.0069191112286514705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,10240,2048,0.02602844436963399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,7168,0.014144000079896716
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,128,5120,0.00888533310757743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2560,2048,0.004319111092223061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,4096,3072,0.007618666523032718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,7168,0.012620444099108378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,12288,0.23173954751756456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,12288,0.0487697786755032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,12288,0.03513866662979126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,512,0.0035084444615576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,256,0.011341333389282227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,5120,0.016896888613700867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,2048,0.04970577690336439
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,128,0.015875554747051664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,1536,0.018209778600268893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,65536,5120,0.1544017791748047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,256,0.002206222257680363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,768,0.011391110718250275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,512,0.002696000039577484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,12288,32,0.0028577778074476454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,1024,0.0058400001790788435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,2560,0.04500000013245476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,3072,0.013120888835854001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,128,0.005606222069925732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,1536,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,10240,0.022587555978033278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,768,0.007251555720965068
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1024,5120,0.008640000389681922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,10240,0.012067555553383298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,65536,0.07955111397637261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,768,0.017158221867349412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,2048,0.042083554797702365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,512,7168,0.010016889207892949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,256,0.01423733267519209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,6144,5120,0.018769777483410306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,2560,0.008636444807052612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,1536,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,6144,0.05513422356711494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,2560,1536,0.045216000742382474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1536,256,0.01517688896920946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,16384,0.0059279998143513995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,512,128,0.004019555532270008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,128,0.014787554740905762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,3584,0.0075333333677715724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,10240,0.05046666661898295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,6144,0.01071200023094813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,128,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,3072,0.023729777998394434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,8192,0.007674666742483775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,512,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3584,256,0.0023822221491071912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,10240,0.047750221358405225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,3584,0.02569955587387085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,256,256,0.005711110929648082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,1024,0.006296000132958095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,3584,0.02850666642189026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,12288,0.06650488906436496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,768,0.006962666908899943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,128,0.014626666903495789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,128,6144,0.01205244412024816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,2560,0.004944889081848992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,1024,0.0031786666562159858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,2048,0.00424888895617591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,512,16384,0.0287715560860104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,6144,0.04582311047448052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2048,3584,0.0099653336736891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,1536,0.006752000086837345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,10240,0.2281573348575168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,3584,0.15542577372656927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,2560,0.011984888878133563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,1536,0.012053333222866058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,8192,0.038732445902294584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,512,5120,0.01038577738735411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,6144,2048,0.022680888573328655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,6144,3072,0.007339555356237624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,768,0.017630222770902846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,2560,0.007286222444640265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,768,0.011335999601417117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3584,2048,0.020583111378881667
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,12288,0.01716088917520311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,12288,128,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,2560,0.0069360000391801195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,4096,0.009721777505344814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,5120,0.023192000057962205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,2048,0.012070222033394707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,12288,1024,0.019996444384257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,7168,0.011393778026103973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,128,0.0181377778450648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,3584,0.016895999511082966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,6144,0.07430044147703382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,1536,0.004467555632193883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,16384,0.01941688855489095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,65536,0.14681688944498697
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,7168,0.0196088883611891
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,128,0.016175111134847004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,3072,0.007432888779375289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1024,2560,0.010787555740939246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,65536,512,0.05674311187532213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,12288,0.049157334698571094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,7168,256,0.002630222174856398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,32,8192,0.010657777388890585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,8192,0.0041751112374994485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,7168,0.004545777622196409
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,2048,0.014470222923490735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,16384,10240,0.07329244746102227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,64,0.002135111019015312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,2048,0.004517333375083076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,256,0.014811555544535318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,5120,0.03063911199569702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,4096,0.054456889629364014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,10240,0.03994399971432156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,1024,0.03572444452179803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,5120,0.03566666775279575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,7168,0.11520089043511285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,3072,0.007330666813585494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,7168,0.009692444569534725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,3584,0.09608711136711968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,5120,768,0.01685333251953125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,16384,0.016413332687483895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,4096,6144,0.026887110537952844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,5120,0.006966222491529252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,12288,0.026783999469545152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,3584,0.007304888632562425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1536,1024,0.014984889162911309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,7168,16384,0.07168710894054837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,1024,0.009458666874302758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,32,0.011519110865063138
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,5120,0.02497422198454539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,7168,32,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,8192,0.05715822511249118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,65536,0.12867289119296604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,8192,0.014396443963050842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2560,3584,0.008287111090289222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,1024,0.014807111687130399
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,3584,0.015613334046469795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,32,0.002185777864522404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,5120,256,0.0031662223239739737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,512,0.02053866618209415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,12288,4096,0.030272000365787085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,768,0.006715555571847492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,65536,0.03085955646302965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,128,0.008251555263996124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,3584,0.01202400028705597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,5120,0.06841599941253662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,768,0.025031111306614343
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2560,10240,0.03312177790535821
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,4096,0.010684444672531553
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,12288,0.029792000850041706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,256,0.006272000157170826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,5120,0.0329386658138699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,7168,0.04407911168204414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,6144,256,0.016505777835845947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,32,1024,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,32,0.0018026667336622875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,2560,2048,0.01685688893000285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,8192,0.08247022496329413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,768,0.003552000141806073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,128,6144,0.008995555341243744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,7168,12288,0.01908888916174571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,128,32,0.0018168888572189545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,12288,12288,0.05485511157247755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,6144,0.015154666370815702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,12288,0.018940443793932598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3072,1536,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,12288,5120,0.03959733247756958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,2560,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,64,0.0022897778285874263
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,5120,0.012738666600651212
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,65536,0.06249511241912842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,2048,0.031748443841934204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,5120,0.02552266584502326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,5120,0.07247555255889893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,8192,0.011015111373530494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,768,0.005997333261701796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,8192,0.03770044445991516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,5120,0.06521066692140368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,5120,0.009662222531106737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,768,2048,0.00981422182586458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,7168,0.04649688800175985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,12288,0.01721155477894677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,4096,0.025683555338117812
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,4096,0.007629333270920648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,8192,6144,0.02884711159600152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,65536,0.02419022222359975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,65536,0.26599733034769696
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,1536,0.004268444246715969
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,2560,0.027123555541038513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,4096,0.008960000342792934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,6144,1536,0.009579555855857002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,3584,0.007643555601437886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,512,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,65536,0.07146222061581083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,12288,0.030469334787792627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1024,8192,0.010671999719407825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,2560,0.01443822185198466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2048,12288,0.013965333501497904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2048,8192,0.03125777840614319
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,8192,3584,0.024937777055634394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,7168,0.03243733445803324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,5120,0.02054755555258857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,1536,0.053391999668545194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,3072,0.05585155884424845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,8192,0.024600888291994732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,12288,0.037238223685158625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,512,0.003532444437344869
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,128,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,5120,0.024829333027203877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,16384,0.05160977774196201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,128,0.0021262222694026097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,12288,0.015136889285511441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,4096,0.013945778210957846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,2048,0.018569777409235638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,3072,0.02514311174551646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,4096,0.009290667043791877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,32,0.00591822216908137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,512,0.006958222223652734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,7168,0.027501333090994094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,128,10240,0.01202666676706738
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,1536,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,5120,256,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,6144,0.11234845055474176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,6144,0.013816888961527081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,64,0.0018542222678661346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,128,0.014486221803559197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,256,0.002187555655837059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,5120,0.013287111288971372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,2048,5120,0.009024000002278222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,6144,0.029253333806991577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,16384,0.022153778208626643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,256,0.006783111227883234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,256,0.01221333361334271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,65536,0.014708444476127625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,1024,0.0035075553589397003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,768,0.005327111317051782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,128,0.01442311041884952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,5120,1536,0.004556444370084339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,3072,0.015271999769740634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,3584,3584,0.007699555820888943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,6144,0.010249777800507015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,2560,0.024345777100986902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,6144,0.009575111170609793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,65536,16384,0.17608533965216744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,8192,0.02887111239963108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,10240,0.017103110750516255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3072,8192,0.02573866645495097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,256,0.015112888481881885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,4096,0.014746665954589844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,64,0.005532444351249271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,3072,768,0.02032711108525594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,10240,0.034662223524517484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,6144,0.03491466575198703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,2560,0.03271289004219903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,10240,512,0.0064035554726918536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,8192,0.015778667396969266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,8192,0.020558221472634208
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,256,0.00544177782204416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,3072,0.05980177720387777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,3584,0.028103109863069322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,4096,0.005618666609128316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,768,0.013760888742076026
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,5120,0.0310204459561242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2560,256,0.00527022240890397
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,4096,0.007918222082985772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,16384,256,0.004096888833575779
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,2560,0.012705778082211813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,6144,0.045363555351893105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,2560,0.007642666498819987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,1024,65536,0.08055466413497925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,6144,256,0.005895111295912001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,128,0.016508445143699646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,65536,0.08270844486024645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,7168,2560,0.012122666670216454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,768,0.0031804444475306403
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,5120,0.009416888985368941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,10240,0.02564711040920681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,6144,0.008959111240175035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,2048,0.006072000082996156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,512,0.010687111152542962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,5120,0.14534666803148058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,3584,0.01479555500878228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,64,0.0069413334131240845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,512,0.003195555466744635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,5120,0.0038604442444112566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,5120,0.049880888726976186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,128,0.005939555664857228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,256,0.0021377777059872947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,1536,0.009230222139093611
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,8192,0.0255057778623369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,256,0.012396444049146442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,65536,0.5286746554904515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,2048,0.006484444356626934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,128,2048,0.006058666855096817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,3072,0.007621333003044128
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,512,0.00693866651919153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,10240,3072,0.01240355521440506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,256,0.005599110904667113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,2560,0.005552000055710475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,128,0.0019422223170598347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2048,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,4096,0.005965333431959152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2560,512,0.003170666595300039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,7168,0.015056000815497505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,10240,0.2412275473276774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,8192,3072,0.02737955583466424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,8192,5120,0.014695111248228284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,16384,0.006632888896597757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3072,768,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,5120,3584,0.027115555273161993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,6144,3584,0.007997333175606197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,64,0.0025395556456512874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,65536,0.11438932683732773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,65536,0.3746062119801839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,1536,0.007244444555706448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,8192,0.023434667123688593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,12288,0.05029866761631436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,6144,0.01682844426896837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,10240,0.0144533332851198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,7168,0.022449778185950384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,8192,0.01310222182008955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,128,6144,0.018188445104493033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,7168,0.019523556033770244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,65536,0.15466845035552979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,2048,0.006623999940024481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,256,0.00638488887084855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,64,0.0019342222561438878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,12288,3584,0.015331554743978711
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,65536,0.41404088338216144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,65536,0.0628897746404012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,2560,10240,0.027671999401516382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,10240,0.04669777883423699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,6144,0.01690933273898231
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,256,8192,0.03171111146608988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,256,0.002463111033042272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,12288,5120,0.03826399975352817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,4096,2560,0.006948444578382704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,32,0.008316444853941599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,16384,768,0.026995556222067937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,12288,2048,0.028168890211317275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,10240,0.0698417756292555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,3072,0.010840000377760993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,6144,2560,0.023610666394233704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,8192,0.02940888868437873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,32,0.001822222231162919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1536,6144,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,768,7168,0.0258559998538759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,2048,4096,0.008007999923494127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,1024,0.015594666202863058
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,512,0.011700444751315646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,16384,0.03257333238919576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,256,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,5120,0.023637332850032385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,16384,0.02548355526394314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,256,0.02087644404835171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,16384,0.28079644838968915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,65536,10240,0.11279200183020698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,768,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,7168,0.008288000192907121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,3584,10240,0.04684355523851183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,1536,0.00794222205877304
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,6144,0.017901332841979135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,5120,7168,0.012187555432319641
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,1536,0.009002666506502364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,6144,0.1273804505666097
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,8192,0.012422222230169507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,16384,10240,0.03437155485153198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,16384,0.0273964438173506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,3072,0.07274222373962402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,12288,0.02497422198454539
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,5120,0.008633777499198914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,10240,0.015375110838148328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,512,0.0024666666156715816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,65536,7168,0.16331999831729466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,7168,0.03529777791765001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,1024,0.030806223551432293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,6144,0.019351111518012155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,65536,0.1449191172917684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,10240,0.03257066673702664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,3072,0.018940443793932598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,64,0.0021955555097924336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,3072,0.005223999834722943
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,4096,0.09540888998243545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,256,0.008634666601816813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,2048,0.004872000051869286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,768,0.002537777854336633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,32,128,0.0017964444640609955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,4096,5120,0.1614168882369995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,65536,0.0793164438671536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,64,0.005667555663320754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,10240,0.04346399837070041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,3584,0.004179555508825514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,12288,128,0.01481244464715322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,3072,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,2560,0.0074417781498697065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,5120,768,0.023807111713621352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,5120,768,0.01682844426896837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,3584,0.006237333433495627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,12288,0.057242665025922984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,768,0.015853333804342482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,256,0.005209777918126848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,512,768,0.003168888803985384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3072,1536,0.006628444625271692
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,768,0.00554577757914861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,128,0.00794666674402025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,128,2560,0.00979555563794242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,8192,2560,0.011406222151385413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,256,0.005589333259397083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,256,0.0021804444905784395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,64,1024,0.0031662223239739737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,256,0.012392889294359418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1024,5120,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,7168,0.01884977850649092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,2560,0.006917333437336816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,12288,0.01719555589887831
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,8192,0.11566132969326443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,256,0.005218666460778978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,768,0.05492888887723287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,10240,0.012430222498046027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,3584,0.012379555238617791
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,8192,768,0.0063662222690052455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,768,0.005592888842026393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,768,0.019241778386963736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,3584,0.061724444230397545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,64,0.00527555536892679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,7168,0.019715555840068393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,10240,0.022341334157519873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,768,0.0062871111763848206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,7168,8192,0.017521777086787753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,256,1024,0.0032115555885765287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,32,10240,0.026945778065257605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,64,0.005316444569163852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,2048,65536,0.10463911294937134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,5120,0.014816888504558139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,16384,0.04046222236421373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,64,0.005375111268626318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,3584,0.040884445110956825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,4096,0.007962666451931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,512,0.022665777140193518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,512,8192,0.011063999599880643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,10240,64,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,7168,10240,0.027090667022599116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,5120,16384,0.0599457754029168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,2560,0.025569778349664476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,16384,0.05705955293443468
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,4096,5120,0.02442311081621382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,256,0.007988444632954067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,512,0.03987200061480204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,4096,0.008314666648705801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,16384,0.02036888897418976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,8192,0.5834710862901475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,16384,0.021260443660948012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,512,0.005888888819350137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,10240,0.014497778481907315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,7168,0.05512266688876682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,256,0.002191999927163124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,256,0.007456888755162557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,3584,0.0681413345866733
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,7168,10240,0.03459377752410041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,2560,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,768,0.0052879999081293745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,65536,0.06743911239835951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,768,512,0.0028408887899584244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,2560,0.01621955633163452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,1536,0.013767110804716745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,32,1536,0.010103999740547603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,1024,0.01519555515713162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,5120,0.013088888592190213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,7168,0.016506666938463848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,10240,768,0.007126222054163615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,768,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,65536,2048,0.23921065860324434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,6144,0.033187554942237005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,2560,0.011228444675604502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2048,768,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,2048,65536,0.1666897800233629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,8192,0.12818400065104166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,2048,0.004885333279768626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,128,8192,0.005923555543025334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,3584,0.032771554258134626
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,2048,0.004912000149488449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,8192,0.012857777376969656
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,1536,0.006616000086069107
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,256,0.03607911202642653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,3584,32,0.005969777703285217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3072,64,0.021295999487241108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,16384,10240,0.05462933248943753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,2560,0.03634666734271579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,8192,10240,0.019559999306996662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,512,0.007307555940416124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,1536,0.006385777973466449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,3072,32,0.0075937774446275495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,32,0.005569777968857024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,1024,0.016883555385801528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,2048,0.007109333243634965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,3584,0.008302222523424361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,10240,8192,0.02270222206910451
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,4096,0.21033777130974662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,6144,0.01131377783086565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,256,0.002587555597225825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,4096,0.015242666006088257
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,768,0.007282666862010956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,1024,0.014874666929244995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,2048,0.004223111189073986
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,4096,0.052687111828062266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,12288,5120,0.016161777906947665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2560,12288,0.09009510940975612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,1024,0.06491377618577746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,2560,5120,0.00962488849957784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,3072,0.009700444837411245
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,1536,0.3850746684604221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,4096,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,4096,0.007983110845088959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,1024,0.007281777759393056
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,16384,0.5442959997389052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,8192,16384,0.025721778472264607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,2048,16384,0.02163022259871165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,16384,0.01621955633163452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,12288,1536,0.010391111175219217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,256,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,3584,0.04076355695724487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,2048,0.006613333192136552
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,6144,0.01717511150572035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,1024,0.004572444491916233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,4096,0.34672265582614475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,10240,0.016460445192125108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1024,2048,0.006244444598754247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,6144,4096,0.027841776609420776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,512,0.00553688903649648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,3072,0.00834222220712238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,16384,0.01682222220632765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,3584,0.006218666831652324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,128,768,0.0025173332542181015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,3584,0.006680888848172293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,8192,0.019985778464211356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,4096,8192,0.013438222308953604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3584,10240,0.03285155693689982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,3072,0.07695021894243029
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,6144,0.009688888986905416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,256,256,0.005676444619894028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,1024,0.0038728887836138406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,32,0.006604444649484422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,12288,12288,0.06518844763437907
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,3072,6144,0.010328888893127441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,1536,1536,0.006228444476922353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,10240,0.027873777680926855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,7168,0.030080888006422255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,16384,6144,0.024824000067181055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,2560,0.015853333804342482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,128,65536,0.025253333979182776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,768,0.006310222049554189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,768,256,0.006266666783226862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,128,0.007296000089910295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,6144,65536,0.07446844710244073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,1536,0.019426667028003268
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,4096,8192,0.0151582227812873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,512,0.012905778156386482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,12288,0.017383999294704862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,32,0.008316444853941599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,256,6144,0.012296888563368054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,1024,768,0.005586666779385672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,16384,0.016202666693263583
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,7168,0.017356443736288283
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,4096,0.036438223388459944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,128,0.015125332607163323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,2048,0.007259555160999298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,4096,0.007707555260923173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,256,0.01884622209601932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,768,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,16384,16384,0.043465776575936206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,768,128,0.005199999858935674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,4096,6144,0.012278222375445895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,768,256,0.0052879999081293745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,2560,0.049661331706576876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,12288,0.0255386663807763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,3072,0.03600000010596381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,8192,0.07362844546635945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,1536,0.00830666638082928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,1024,5120,0.019225777851210702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,1024,0.011725333001878528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,128,0.005558222118351195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,65536,0.04692088895373874
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,256,32,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,5120,0.021553777986102637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,3072,0.06597866614659627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,1024,0.003160888950030009
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,512,8192,0.010693333215183683
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,3072,0.05084355672200521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,12288,12288,0.05024266574117872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1536,10240,0.012404444317022959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,512,0.012723555167516073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,32,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,3584,0.10329688919915093
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,256,0.002834666727317704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,128,0.005571555760171678
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,10240,512,0.006248888870080312
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,1536,0.03660800059636434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,6144,0.007720000214046902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,5120,0.01573155489232805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2048,512,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,128,1536,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,64,3584,0.004912888838185204
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,1536,256,0.00216977774269051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,8192,0.014811555544535318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,32,64,0.006216889040337668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,768,0.009313777916961247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,2560,0.011728888584507836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,1536,256,0.006959111326270633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,2048,0.00646311127477222
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,2560,0.007526222202512953
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,5120,3072,0.168013334274292
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,7168,0.02938488788074917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,6144,0.013665777941544851
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,3584,0.00832177781396442
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,4096,0.04820000131924947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,1024,1536,0.0038355555799272326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,1536,0.020596444606781006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,6144,0.030814223819308813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,12288,0.03503288825352987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,768,1536,0.014078223043017916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,768,0.005620444400442972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,12288,256,0.01722400055991279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,32,0.00180177783800496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,12288,0.013408888545301227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,128,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,3072,0.007263999846246507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,12288,0.016524444023768108
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,2560,0.007640889121426477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,6144,0.01716977854569753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,2048,0.008958222137557136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,6144,0.00998044427898195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,4096,32,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,8192,0.01035911093155543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,768,0.024367999699380662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,7168,0.0410480002562205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,5120,0.02076088885466258
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,768,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,256,0.00464088883664873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,5120,0.008952000074916417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,7168,0.022943110929595098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,512,0.012749333348539142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,768,0.002924444360865487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,4096,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,12288,0.04579022195604113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,12288,0.02017155620786879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,6144,0.02476177778508928
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,32,0.0021537778278191886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,4096,0.035938666926489934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,3072,0.006911110960774952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,2560,0.007633777956167857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,128,0.0024871110088295406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,7168,0.010028444230556488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,16384,0.01332266628742218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,4096,256,0.01514755520555708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,3584,0.007288888924651676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,7168,0.0601102246178521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,512,4096,0.01182577759027481
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,5120,0.009400889277458191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,128,0.001810666587617662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,2560,0.007290666302045186
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,512,0.016196444630622864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,768,16384,0.02164177762137519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,2048,0.023405333360036213
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,256,12288,0.010244444840484196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,10240,0.020432889461517334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,128,3584,0.014072888427310519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,6144,0.024711110525661047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,512,768,0.01564800077014499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,65536,0.6854169103834363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,12288,0.01900888813866509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,2048,0.009646221995353699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,256,7168,0.007236444287829929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,3584,0.017490666773584154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,128,1024,0.008658666577604082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,768,128,0.003216888962520493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,512,65536,0.07038844294018216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,768,0.0041831110914548235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,12288,0.04986844460169474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,128,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,16384,0.20833244588640001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,3584,512,0.005951111101441913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,10240,0.015458666616015963
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,32,0.0018399999373488957
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,1024,0.003141333245568805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3584,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,16384,3072,0.02735822233888838
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,1024,1024,0.013157332936922709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,1024,0.00999644481473499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,512,16384,0.16008888350592718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,1024,0.018544889158672757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,1024,0.004334222111437055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,2048,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,65536,1024,0.03345511025852627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,2560,0.016349333855840895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,4096,6144,0.012759110993809171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,8192,0.009336888790130615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,3584,0.103657775455051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,2560,0.012088889049159156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,16384,8192,0.07845066653357612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,5120,0.06949600246217516
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2048,6144,0.009659555223253038
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,4096,0.025148444705539282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1024,256,0.00249777775671747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,256,0.015475556254386902
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,10240,65536,0.14386933379703098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,256,10240,0.03855289022127787
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,4096,0.04945688777499729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,8192,0.021314667330847845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,1024,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,768,0.017895999881956313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,768,0.013744889034165276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,8192,7168,0.03550844391187032
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,256,0.020282665888468426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,5120,0.008394666843944127
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,1024,0.021966222259733412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,2560,0.024337778488794964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,65536,0.11385067303975423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3072,128,0.014492443866199918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,768,0.014781332678265043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,2560,0.008311111066076491
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,256,3072,0.00795733349190818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,5120,65536,0.07532444265153673
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,10240,5120,0.01882844501071506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,64,0.0019617778145604664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,65536,0.2983235518137614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3072,5120,0.01090755561987559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,256,0.0018773333479960759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,8192,12288,0.04100088940726386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,3584,0.04302311274740431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,2048,0.008629333641793992
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,1024,0.022513777017593384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,2560,512,0.014402666025691561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,5120,0.014488000008794995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,768,3584,0.07435466845830281
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,5120,0.1668435600068834
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,10240,0.04155022237035964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,10240,0.018004443910386827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,5120,256,0.015452444553375244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,16384,0.02296000056796604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,8192,0.03990044527583652
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,5120,1024,0.006262222097979651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,16384,0.5205350981818305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,6144,8192,0.01679377754529317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,4096,0.03333866596221924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,1536,0.0036595554815398324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,16384,0.06014755699369642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2048,12288,0.016306666864289176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,10240,0.04594044552909004
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,65536,2560,0.028361777464548748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,16384,256,0.017097777790493436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,512,0.013968000809351603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,10240,0.02438666588730282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,768,0.014422222971916199
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,10240,0.16368267271253797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,2048,4096,0.013119999733236102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,256,0.014146667387750415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,65536,0.06325066751903959
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,10240,4096,0.015838222371207345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,65536,7168,0.1618728902604845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,16384,6144,0.024969778127140466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,2560,0.007003555695215861
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,16384,0.015812445018026564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,3072,0.008620444271299574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,128,0.0031831111345026228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,256,4096,0.01568622224860721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,64,0.0020231112009949153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,3584,0.019867555962668527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,4096,65536,1.5357066260443792
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3072,16384,0.023372444841596816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,7168,0.018230222993426852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,10240,0.020587555236286588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,512,2048,0.014095111025704278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,2048,0.02142933342191908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3072,128,0.008608000146018134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,7168,0.010048000348938836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,32,0.005568000177542369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,4096,0.008072888685597314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,1536,0.028408888313505385
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2560,7168,0.011736888852384357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,12288,0.0455040004518297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,256,0.039460443788104586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,12288,7168,0.02017600006527371
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,32,0.0018213333355055915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,512,0.006618666566080517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,10240,0.19015199608272979
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,3072,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,4096,0.01310488912794325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,768,3584,0.0075502221783002215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,1536,256,0.005199999858935674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,32,0.0018257778137922287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,256,0.01685688893000285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,10240,0.049068444304996066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,32,0.00517155561182234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2560,12288,0.014484445254007975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,64,128,0.006364444477690591
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,768,0.005931555396980709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,4096,128,0.005893333504597346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,5120,0.009119999905427298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,512,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,2048,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,768,0.02937600016593933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,128,0.004179555508825514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,5120,0.0041857775714662345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2560,6144,0.0685031082895067
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,512,3584,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,8192,0.017103999853134155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,3072,0.010816888676749336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,8192,256,0.007073777417341868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,2048,0.004571555389298333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,3584,12288,0.017886221408843994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,4096,0.015292444162898593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,10240,0.05729866690105862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1024,1536,0.0038684445122877755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2048,7168,0.08567644490136041
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,8192,16384,0.0529075562953949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,3072,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,16384,7168,0.04439199964205424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,32,5120,0.005936000082227919
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,7168,64,0.0018595556418100994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,128,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,4096,0.07864533530341254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,7168,0.018055111169815063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,12288,0.028350220786200628
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,512,12288,0.04232711262173123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,12288,3072,0.01499111122555203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,512,768,0.005559111220969095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,8192,0.013082666529549493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,3072,0.020615999897321064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,7168,0.013032000097963544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,4096,0.016336888074874878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1536,16384,0.03149244520399306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,512,0.012716444830099741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,7168,64,0.0025457777082920074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,4096,10240,0.014939554863505893
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,128,2048,0.008671999805503422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,12288,0.01919911139541202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,12288,7168,0.03656889001528422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,3072,1536,0.016206221448050607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,7168,1024,0.01788444485929277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,128,0.002161777681774563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,768,256,0.005882666756709416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,768,256,0.005598222215970357
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,2560,5120,0.009464888936943477
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,256,768,0.005443555613358815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,2560,0.04878133204248217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,128,768,0.018212444252438016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,4096,5120,0.011235555840863122
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,5120,0.015461333923869662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,128,1024,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,1024,0.05134933193524679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,10240,32,0.00665244460105896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,5120,0.04309333364168803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,5120,8192,0.07621866464614868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,128,0.014087999860445658
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,65536,0.059699555238087974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,7168,0.01127466642194324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,7168,12288,0.0410097771220737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1536,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,128,0.014468444718254937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,64,0.0025439999169773525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,5120,6144,0.013174222575293647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,7168,0.0144533332851198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,128,10240,0.012055111428101858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,768,768,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2048,64,0.009370666411187913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2048,65536,0.20535643895467123
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,4096,0.009165333376990424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,10240,0.02125866711139679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,768,0.007081777685218387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,6144,65536,0.08582044310039944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,8192,0.9167795181274414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,10240,0.04675999946064419
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,128,2048,0.014464888307783337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1024,4096,0.013800000150998434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,1024,0.02366311185889774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,16384,0.04382044408056471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,5120,0.024711110525661047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,6144,2560,0.02540622154871623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,256,0.0029537777105967202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,4096,0.00796888851457172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,6144,0.004559111264016894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,1536,0.009192888935407003
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,16384,5120,0.01754755609565311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,7168,3072,0.010659555594126383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,10240,0.3498488797081842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,5120,0.01961599952644772
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,3584,0.07898310820261638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,5120,0.023426666855812073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,1536,6144,0.018572444717089336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,256,0.0058746664888328975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,5120,0.007027555671003129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,32,0.0069822221994400024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,64,0.0017973333597183228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,512,0.005214222189452913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,16384,0.013878222141000958
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,3584,0.07332533597946167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,12288,0.014476444986131458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,16384,4096,0.02271288964483473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,1536,0.027671999401516382
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,3584,128,0.014748444159825643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,7168,0.012448888685968188
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,4096,6144,0.01924800044960446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,16384,128,0.00686488880051507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2560,128,0.005946666830115848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,12288,0.07243911425272624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,5120,6144,0.014922666880819531
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,128,32,0.0018195555441909367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,1024,0.006302222195598815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,10240,0.015801777442296345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,1536,0.006600889066855113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2048,2560,0.018446221947669983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1536,3584,0.03307644526163737
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,512,12288,0.024383111132515803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,3072,0.09541510873370701
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,12288,0.013809777796268463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3584,16384,0.02293066680431366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,8192,3072,0.027828445037206013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,768,0.016395555602179635
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,65536,4096,0.8666870858934191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,256,0.00318755561278926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,512,0.010695111420419482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,7168,0.02401777770784166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,128,0.0028115556471877625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,1536,0.022608000371191237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,8192,0.011184000306659274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,512,0.012808889150619507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,768,2048,0.0042568888101312844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,3072,0.007276444799370236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,2048,0.00572533326016532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,3072,0.03681066632270813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,32,6144,0.006125333408514659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,7168,256,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,16384,3072,0.013767999907334646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,65536,0.061272886064317494
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,10240,0.04920711119969686
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,6144,0.009735999835862054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,1536,0.007989333735571967
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,7168,1536,0.020250666472646926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,2048,0.01479466590616438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,12288,0.01591199967596266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,768,0.005952000204059813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,128,12288,0.021712889273961384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,2560,0.026709333062171936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,10240,0.05383822321891785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,2560,0.05991466840108236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,8192,0.03920266694492764
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,65536,0.05074310965008206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,128,1024,0.009518221848540837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,6144,0.009016888837019602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,5120,0.028704000843895808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,2560,0.2016026708814833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,10240,512,0.007975111405054728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,64,3584,0.010703999963071613
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,256,3072,0.013766222529941134
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,7168,0.01145155562294854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2048,6144,0.010603555374675326
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,5120,256,0.005594666633341048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,32,0.002598222138153182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3072,512,0.015636444091796875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,2560,0.008959111240175035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,2560,0.08005066712697347
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,128,0.006632888896597757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,3072,0.007322666545708974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,3584,2560,0.008727110922336578
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1024,2560,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,3584,0.007836444510353936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,512,0.03383822242418925
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,8192,16384,0.02719377809100681
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,3072,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2560,5120,0.00928177767329746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,1024,0.013447110851605734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,768,7168,0.010038221875826517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,16384,0.02215911116864946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,64,4096,0.007648000286685095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3584,512,0.002814222127199173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,8192,0.017010667257838778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,512,0.002480000050531493
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,1024,5120,0.008332444561852349
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,1536,0.008619555168681675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,768,8192,0.020756444997257657
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,1536,0.00628355559375551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,4096,0.00850933293501536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,10240,7168,0.02054222259256575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,2048,0.01648977730009291
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,256,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,6144,0.015473778049151102
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,10240,3584,0.030221333106358845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,3072,0.032419555717044406
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,5120,0.07591822412278917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,16384,0.021889777647124395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,256,2560,0.014434667097197639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3584,65536,0.06767733229531182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,4096,0.039763556586371526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,128,1536,0.0351528889603085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,5120,256,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,768,0.0028515555378463534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,3584,0.08079822195900811
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,16384,0.04124177826775445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,5120,10240,0.0273324449857076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,7168,32,0.0021520000365045336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,65536,0.12607022126515707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,65536,0.06688888867696126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2048,12288,0.0120000003112687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,5120,0.008438222110271454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,10240,0.012060444388124677
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,2048,0.0042275554604000514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1536,7168,0.013411555853154926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,5120,5120,0.013065777719020844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,12288,1.0705689324273002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,3072,0.013739555246300168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2048,2560,0.007266666326257918
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,256,0.007607111500369177
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,256,0.00962488849957784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,4096,6144,0.020807999703619216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,12288,0.12488088342878555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,3584,0.00962755580743154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,16384,0.0162426663769616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,12288,0.018726223044925265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,3072,16384,0.025619554850790236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,2560,256,0.006565333240562015
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,7168,0.15390577581193712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,8192,2048,0.0889128910170661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,32,3072,0.0075804442167282104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,2560,0.01924444403913286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,1536,0.025381333298153345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,6144,0.022707555029127333
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,10240,0.01718755563100179
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,512,128,0.005543111099137201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,8192,0.04092533389727274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,65536,0.09236444367302789
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,2560,3072,0.021479111578729417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,4096,0.011410666836632622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,768,65536,0.2917911211649577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,1024,6144,0.02439555525779724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,12288,0.02776088813940684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,5120,0.028718223174413044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,12288,2560,0.04911288950178358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,2048,0.0069120000633928515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,5120,0.00943377779589759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,512,12288,0.0233608881632487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,3584,0.007920000288221572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,3072,0.03636888994110955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,256,0.015234667393896313
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,4096,16384,0.04471022221777174
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,128,32,0.0069413334131240845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,128,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,4096,0.0086986662613021
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,1536,5120,0.008651555412345463
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,512,0.01716711123784383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,128,64,0.0017964444640609955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,128,256,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,2048,0.06299200322892931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,3072,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,10240,512,0.019280888968043856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,2048,0.008283555507659912
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,8192,0.02163822286658817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,2048,0.023152889476882085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,1024,2560,0.017851556340853374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,4096,12288,0.058931556012895375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,768,6144,0.017848889032999676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,65536,0.17288977570003936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,768,65536,0.02032088902261522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,1536,0.005559999909665849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1536,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3584,5120,0.03052266769938999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,6144,128,0.009650666680600908
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,7168,256,0.005592000153329637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,128,0.015094222293959724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,16384,0.004931555440028508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,7168,0.02060622142420875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,1536,0.26774578624301487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,65536,0.07381777630911933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,5120,12288,0.029734220769670274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,5120,3584,0.007797333101431529
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,5120,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,10240,0.03251555562019348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,16384,0.010997333460383944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,256,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,8192,0.014358222484588623
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,8192,1536,0.019729778170585632
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,1024,0.006257777826653586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,32,2560,0.006953777538405524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,768,768,0.0028133332315418455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,12288,0.12696088684929743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,1024,0.00590133335855272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,6144,64,0.011211555865075854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,2048,1536,0.0069546666410234236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,512,0.08207822508282132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2560,32,0.0026142222599850762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,8192,0.02572711143228743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,1024,0.010471999645233154
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,12288,0.053800000084771045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,1024,0.0028293333533737394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,10240,0.1591591172748142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,1536,0.005290666802061929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,256,0.009665778113736046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,7168,0.1683431069056193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,16384,16384,0.05094844433996412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1024,1536,0.014152000347773233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,5120,0.008616888688670265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,8192,2560,0.024754666619830664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,12288,0.05257599883609348
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,512,0.00555377784702513
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,8192,64,0.027141332626342773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,1024,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,8192,0.01131199962562985
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,1536,0.006753777878151999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,2560,0.024655999408827886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,1024,0.012061333490742577
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1536,2048,0.0069360000391801195
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3584,128,0.00565244464410676
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,768,10240,0.00869777798652649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,2048,7168,0.008993777963850234
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,65536,256,0.024288889434602525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,256,128,0.005618666609128316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,128,0.015791111522250705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,32,0.005621333503060871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3072,768,0.006959111326270633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,4096,0.011399110986126794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,7168,0.012425777812798819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,2560,0.014366222752465142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,8192,0.03228177626927694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,3072,7168,0.012668444050682915
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,2560,128,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,128,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,10240,1536,0.009328000247478485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,12288,0.011038222246699862
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,32,512,0.002473777780930201
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,3584,0.010314666562610203
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,2048,0.042971554729673594
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,3072,0.007638221813572778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,3584,0.00831733312871721
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,256,0.017508443858888414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,64,0.01038044442733129
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,4096,6144,0.012444444828563266
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,16384,3072,0.26125778092278373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,4096,0.016307555966907077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,6144,3072,0.19980977641211617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,1536,0.004867555366622077
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,7168,10240,0.017904000149832833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,3072,0.017817777064111497
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,32,0.006255110932721033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,2048,0.010679110884666443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,768,0.007676444119877285
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,1024,0.006630222416586346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,12288,0.5622586674160427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,65536,0.1961359977722168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,2048,0.07274133629269071
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,3072,512,0.002833777831660377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,1024,128,0.004932444542646408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,7168,16384,0.19375733534495035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,12288,0.025739555557568867
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,12288,1024,0.013926222920417786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,7168,3072,0.010464888479974536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,7168,0.025905778010686237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,2048,0.012072000238630505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,65536,1024,0.017351999878883362
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,2560,1536,0.01683288812637329
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,2048,0.005532444351249271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,10240,12288,0.06470400094985962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,10240,0.03015466531117757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,4096,128,0.014480888843536377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,10240,16384,0.04134488768047757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,3072,0.012883555557992725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,256,8192,0.01036888857682546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,16384,0.023578666978412207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,768,3584,0.012210666305489011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3584,4096,0.008669333325492011
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1536,16384,0.01953066719902886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,3072,0.02761777738730113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,8192,0.7670728895399305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,512,3072,0.008781332936551835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,8192,0.09344800313313802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,5120,16384,0.0472871098253462
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,512,0.005904888941182031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1536,3584,0.0211217784219318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,10240,0.1858337720235189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,6144,0.02162577708562215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,8192,0.02443466583887736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,2560,2048,0.020277332928445604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,32,2560,0.00999022192425198
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,512,1536,0.01386044422785441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,12288,6144,0.027651555008358423
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,65536,768,0.29373243119981557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,768,0.026774222652117412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,512,0.011695110963450538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,16384,0.05106044477886624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,32,0.0028026666906144884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,512,0.02086044516828325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,6144,128,0.009004444711738164
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,2560,3072,0.00656444455186526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,64,64,0.001581333370672332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,6144,0.028742220666673448
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,2048,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,12288,5120,0.016520000166363187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,768,0.0038622220357259116
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,3072,16384,0.09781510962380303
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,256,768,0.006622222148709827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,16384,0.046317332320743136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,12288,0.040425777435302734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,256,0.03218133250872294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,64,0.006955555743641323
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,8192,10240,0.027440889014138117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,2560,0.05276088913281759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,8192,2048,0.0075866662793689305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,32,0.005370666583379109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,5120,4096,0.029249777396519978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3072,3072,0.0191928893327713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,65536,128,0.020643555455737643
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,5120,0.011406222151385413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,12288,256,0.006261333409282897
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,4096,2048,0.017200888858901132
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,768,0.01612977849112617
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,3072,4096,0.00793599999613232
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,5120,0.008615111311276754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,16384,1536,0.02981688910060459
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,3584,65536,0.1345813274383545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,512,0.0031759999692440033
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,512,0.0069520001610120135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,3584,0.08842400047514175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,2048,0.008667555948098501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,12288,10240,0.026542221506436665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,1536,0.006903999795516332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,8192,0.024821332759327356
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,8192,0.012096000214417776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3584,768,0.005585777676767773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,3072,0.015419556034935845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,3584,0.04376711116896736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,4096,6144,0.023440000083711412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3584,32,0.002471999989615546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,1024,0.006235555642180973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,2048,0.21300089359283447
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,256,1536,0.0035537779331207275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,8192,16384,0.030799110730489094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,2048,0.047036445803112455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2560,1024,0.0034933334423436057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,6144,0.0357893341117435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,3072,0.007866666548781926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,7168,6144,0.015097777048746744
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,12288,768,0.006743999818960826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,4096,128,0.001847111102607515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,512,0.030176000462638006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2048,65536,0.06383466720581055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,7168,256,0.01716711123784383
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,3584,0.007984889050324758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,2048,0.007376000285148621
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,3584,0.016386666231685214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,512,8192,0.013396444420019785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,32,1536,0.0041759999261962045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,2048,0.03000088863902622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,16384,512,0.0069671107663048645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,8192,1024,0.007981333467695449
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,8192,0.09456444448894924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,256,512,0.01998488936159346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,768,7168,0.05094044407208761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,128,2048,0.007905777957704332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,8192,4096,0.011341333389282227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,6144,0.03395288851526048
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,7168,32,0.002479111154874166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,10240,512,0.009810667071077559
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,8192,0.02914488977856106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,3072,0.007298666569921706
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,8192,0.017687110437287223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,3072,0.014409777190950183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,2048,0.006701333241330252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,512,0.011318222516112857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,7168,6144,0.015012444721327888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3072,1024,0.005580444302823808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,6144,0.11961511770884196
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,128,0.005190222213665645
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,5120,7168,0.013624889155228933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,3072,65536,0.0972213347752889
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,6144,5120,0.014170666535695394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,128,0.005399999933110342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,512,0.0025235555238193935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,8192,0.027048889133665297
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,10240,0.013097777962684631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,512,0.0028106667515304354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,3584,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,256,1024,0.0028897778441508612
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,6144,0.17692000336117217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,2048,8192,0.027441778116756018
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,3584,0.10705333285861546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,128,0.01238844460911221
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,12288,0.012507555385430654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,128,0.020917332834667627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,64,64,0.001828444500764211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,4096,0.019484443797005545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,4096,0.007911110917727152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,2048,0.006329777754015393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,65536,0.08077155881457858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,6144,6144,0.018601778480741713
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3072,1536,0.01498222185505761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,32,2048,0.004433777597215441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,4096,5120,0.017636444833543565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,768,0.05999733342064751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,16384,1536,0.012433778080675336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,128,0.002500444443689452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,65536,0.20370754930708143
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,256,0.00286311118139161
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,6144,0.013474666410022311
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,6144,512,0.005562666803598404
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,3584,2048,0.01997599999109904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,6144,3584,0.009308444129096137
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,512,0.008268444074524773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,8192,0.010712888505723743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,16384,4096,0.018090666996108163
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,7168,0.01036622209681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,1536,0.005912888795137405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,256,16384,0.006188444379303191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,5120,0.007272889216740926
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,10240,0.016495111915800307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,768,128,0.005239999956554837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,4096,64,0.0018275556051068837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,16384,2560,0.012747555143303342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,1024,3584,0.022060443957646687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,256,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,8192,0.11709600024753147
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,256,0.014771555860837301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3584,3584,0.010079999764760336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,2560,0.004996444616052839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,128,768,0.0059004442559348206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,1536,0.005362666729423735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,12288,0.08369511365890503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,512,7168,0.009375111096435124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,4096,0.006231110956933763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3072,5120,0.00850311087237464
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,256,512,0.005383111122581694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,4096,0.007941332956155142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,7168,6144,0.03877066572507223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,7168,512,0.005555555638339784
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,768,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,1536,0.008457777400811514
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1024,8192,0.009305777649084726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,2560,0.01000711073478063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,3584,0.007654222349325816
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,3584,128,0.014459555347760519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,8192,0.16501066419813368
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,64,0.005240889059172736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,32,0.012040889097584618
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,64,16384,0.015483554866578845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1024,10240,0.034789333740870156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,12288,8192,0.0392577780617608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,2560,32,0.007875555919276344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,4096,0.018274666534529794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,6144,8192,0.020595555504163105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,256,0.002463111033042272
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,65536,128,0.027099554737408955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,2048,0.006903999795516332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,512,64,0.006923555499977536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,256,0.009648000200589498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,32,64,0.005541333307822545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3072,8192,0.014004444082578024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,256,4096,0.007948444121413762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,1024,0.08434666527642144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,1536,1024,0.019897777173254225
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,10240,16384,0.05866489145490858
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,2048,0.08242577976650663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,10240,0.05628355344136556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,3584,3584,0.008282666405042013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,16384,0.027954667806625366
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3072,16384,0.020785777105225455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,32,2048,0.015592000550693937
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,5120,0.009001777403884465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,16384,0.035014222065607704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,4096,0.015314666761292351
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,128,0.005550222264395819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3072,16384,0.034621334738201566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,4096,0.016140444411171805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,768,6144,0.008969777988062965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3584,3072,0.007666666474607255
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,1536,0.012749333348539142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,1024,0.027829332484139338
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,5120,0.008633777499198914
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,256,128,0.006739555547634761
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,512,0.056398219532436795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,6144,2560,0.011399110986126794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,8192,5120,0.0317795541551378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,4096,3072,0.009912889036867354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,4096,1024,0.015461333923869662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,7168,0.03426133261786567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,6144,0.029021332661310833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,7168,0.016226665841208566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,256,7168,0.016570667425791424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,256,768,0.0052488889131281115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,768,0.005941333456171884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,16384,10240,0.03507466779814826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,10240,3072,0.02626577847533756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,7168,0.0181404451529185
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,5120,0.008664000365469191
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,4096,4096,0.010441777606805166
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,256,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,128,512,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,16384,0.04231022132767571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,128,512,0.005215999980767568
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,65536,64,0.005207111024194294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1024,5120,0.008970666262838576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,4096,4096,0.029451555675930444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,65536,0.06325333648257785
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2048,3584,0.012068444656001197
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,10240,0.04903911219702827
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,3584,1536,0.016522667474216886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,32,0.001628444426589542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,2048,0.043262223402659096
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,10240,5120,0.5384648640950521
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,3584,65536,2.2082106272379556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,6144,3584,0.012888000243239932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,6144,1536,0.028604444530275133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,32,0.012109333442317115
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,3584,0.007618666523032718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,5120,16384,0.036650667587916054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1024,6144,0.018654222289721172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,512,7168,0.020950223008791607
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,256,8192,0.030292444758945044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,1024,0.02972088919745551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,256,64,0.005897777775923411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,2560,0.0069777775141927935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,2048,0.004222222086456087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,6144,0.0391884446144104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,16384,8192,0.03621422251065572
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,7168,0.015081778168678284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,5120,10240,0.029382222228580054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,4096,0.018925334016482036
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,12288,512,0.02017866737312741
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,512,0.005609777652555042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,10240,768,0.008721777962313758
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,16384,768,0.018921777606010437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,12288,2560,0.019879110985332064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,1024,0.026710222164789837
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,1024,16384,0.01647555496957567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,2048,0.020271999968422782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,5120,8192,0.018374222848150466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,32,0.007265778051482306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,6144,0.10117333465152317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,768,0.02453244394726223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,10240,128,0.00657155571712388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,256,2560,0.006610666712125142
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,256,256,0.0021822222818930945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,3584,10240,0.016318221886952717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,3072,2048,0.006055999961164262
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,2048,0.006681777950790193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,8192,0.01331822243001726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,3584,0.018578666779730055
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,2560,0.008616000413894653
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,2048,0.01424088908566369
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2048,3584,0.024180443750487432
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,8192,16384,0.06295022037294176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,10240,0.012320000264379712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,4096,128,0.0058755555914507965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,16384,6144,0.023419555690553453
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,128,1536,0.0042088888585567474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,768,10240,0.03639555639690823
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,128,128,0.010677333507272931
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,256,1024,0.016525333126386006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,1536,0.007307555940416124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,1536,0.0035413333939181436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,1536,256,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1536,128,0.004918222212129169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,1024,0.006279110908508301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,4096,7168,0.01310755560795466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,1024,0.005607111172543631
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,5120,0.03131466772821214
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,1536,0.012093333734406365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,16384,2560,0.012873777912722694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3584,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,1536,0.006537777682145436
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,64,2048,0.0069422221018208405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,4096,512,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,256,0.015117333994971381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,2048,0.00490133340160052
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,6144,0.01038933296998342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,512,3584,0.015829333000712924
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,1536,2560,0.007794666621420119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,256,0.006647999915811751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,256,0.04407466782463921
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,65536,8192,0.3088337845272488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,10240,0.03208977646297879
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,128,0.0025031111306614345
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,65536,16384,0.7905413309733073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,128,256,0.015082667271296183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,256,0.006263999889294307
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,128,0.035939554373423256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,8192,6144,0.021191999316215515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3072,16384,0.020960888928837244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,768,0.011388444238238864
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,8192,0.3518693447113037
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,256,3584,0.014843554960356818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,4096,5120,0.010302222437328762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,8192,0.005392000079154968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,10240,0.030181331766976252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,65536,768,0.016159110599093966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,65536,5120,0.8700853453742133
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,65536,2.9188550313313804
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,512,0.012105777859687805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,8192,0.021635555558734473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,256,16384,0.04404088854789734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,7168,128,0.007992000215583378
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,256,7168,0.007669332954618666
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,256,768,0.021956443786621094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,2560,0.007257777783605788
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,512,8192,0.04584533307287428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,65536,5120,0.06252977583143446
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,2048,64,0.002854222224818336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,4096,16384,0.27292799949645996
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,4096,2560,0.0069439998931354946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,1536,0.027574222948816087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,5120,0.008971555365456475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,12288,0.02849066588613722
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,4096,0.00793422261873881
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,512,768,0.005995555470387141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,3584,768,0.01852888862291972
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,256,0.00253866674999396
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,12288,1.2074364556206596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,768,1024,0.013785777820481194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,65536,0.06102666589948866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,512,7168,0.005538666827811136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3584,10240,0.015174221661355762
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,1024,2048,0.00867555538813273
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,512,0.007630222373538547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,128,0.005254222287072076
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,6144,8192,0.01938222183121575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,256,0.01796888808409373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,768,0.0069644442862934535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,768,6144,0.00981422182586458
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,128,1024,0.003170666595300039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,2048,0.020555555820465088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,5120,0.006671999891599019
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,65536,10240,0.18836977746751574
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,128,0.013437333206335703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,5120,0.019944889677895438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,10240,7168,0.06547377506891887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,6144,0.009008888569143083
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,10240,3072,0.012087110843923358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,768,0.005761777775155173
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,5120,32,0.002512000087234709
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,64,10240,0.03268977668550279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,64,0.005575110928879843
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,1536,0.018761777215533786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,6144,6144,0.2652542326185438
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,5120,8192,0.03179555469089084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2048,64,0.0052791109515561005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,768,3584,0.007975999679830339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,4096,8192,0.013776889277829064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,10240,0.01919911139541202
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,12288,0.005202666752868229
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,10240,0.24312355783250597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,512,7168,0.010060444474220276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,512,65536,0.10218933555814956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,768,0.007250666618347168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,768,3584,0.017222222354676988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,7168,768,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,768,0.019225777851210702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,768,8192,0.0796639985508389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,768,8192,0.010379555324713389
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,128,0.006574222197135289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,512,0.014835556348164877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,5120,0.06388088729646471
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,5120,12288,0.02678488857216305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,512,8192,0.010599111517270407
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,8192,2048,0.00740799970097012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,2560,256,0.005256888767083486
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,8192,128,0.014771555860837301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,128,10240,0.012829333543777466
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2048,5120,0.01754488878779941
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,1024,0.02657866643534766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,512,0.005209777918126848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,65536,1024,0.06216000186072456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,768,512,0.018173333671357896
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,512,2560,0.015134221977657743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2560,5120,0.011663999822404651
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,1536,0.006618666566080517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,1024,2560,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,1536,0.02364088926050398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,16384,0.012343111137549082
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3584,16384,0.03398133317629496
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,5120,0.009983110758993361
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,12288,768,0.007305777735180325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,512,512,0.002458666761716207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,7168,256,0.015855999456511605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,16384,768,0.014133334159851074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,10240,0.05438933438724942
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,2048,0.009352888498041365
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,768,0.01646933290693495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,768,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,128,0.0018702221827374564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,3584,0.007674666742483775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,32,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,5120,0.02550044490231408
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,256,8192,0.021055999729368422
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,2560,256,0.006590222318967183
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,512,10240,0.01275644451379776
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1024,7168,0.01035822182893753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,10240,0.7548897531297473
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,128,3072,0.008375110725561777
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,5120,0.009327111144860586
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,1024,8192,0.011714666253990598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,768,6144,0.007312888900438945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,4096,0.08025955491595797
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,768,0.03274844421280755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,7168,0.010328888893127441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,16384,768,0.007848000360859765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,6144,0.012461333639091916
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,512,7168,0.019039111004935372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,6144,10240,0.03753866751988729
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,16384,0.005631111148330901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2048,32,0.006575999988449945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,128,12288,0.02221777704026964
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,4096,0.023592889308929443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,1024,7168,0.010384889112578498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,256,0.06169778108596802
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1024,2048,0.005478222337034013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,6144,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,512,16384,0.019946666227446664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,65536,5120,0.04635733366012573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,32,0.006601777755551868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,12288,0.013066666821638743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3584,768,0.00630666646692488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,128,64,0.0018293333964215384
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,5120,5120,0.023271110322740342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,512,0.002518222149875429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,65536,0.08433866500854492
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,3584,768,0.01681422193845113
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,8192,65536,0.08131021923489042
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,7168,0.19822133911980522
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,4096,4096,0.009693333672152625
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,10240,0.04255911045604282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,1536,1024,0.0035093331502543557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,12288,6144,0.12671377923753527
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,3584,0.06293155749638875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2560,10240,0.02546577817863888
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,768,65536,0.06692088974846734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,256,0.005883555445406172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,6144,512,0.0031555555760860443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,32,0.33183378643459743
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,1024,0.01238222254647149
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,10240,0.02786311176088121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,6144,512,0.015825778245925903
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,512,4096,0.005942222144868638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,12288,0.029502222935358684
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,128,2560,0.0069164443347189165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,512,10240,0.021976000732845728
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,12288,0.1885075569152832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,256,7168,0.005551110953092575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,8192,0.014408888088332282
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,128,3072,0.013785777820481194
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,1536,0.03767377800411648
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,10240,16384,0.054528001281950206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1024,2560,0.00488711107108328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,16384,0.029241777128643457
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,16384,0.019780443774329293
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,12288,256,0.002543111021320025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,12288,0.013428444663683573
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,4096,1024,0.006656000183688269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,128,0.002149333349532551
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,768,0.00693866651919153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,6144,0.028732445504930284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2560,768,0.006281777802440856
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,10240,2560,0.012466666599114736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,3584,0.011184000306659274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,65536,512,0.015232889188660515
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,2560,0.011205332974592844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,256,0.0028453332682450614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,2048,0.006287999865081575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,7168,10240,0.036438223388459944
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,256,65536,0.32148265838623047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,1536,64,0.005252444495757421
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,65536,0.6871102121141223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,128,0.005233777893914117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,7168,1024,0.010433778166770935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,128,1536,0.011718221836619906
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,2560,0.009322666459613377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,10240,16384,0.059120886855655246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,768,0.005596444424655702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,7168,16384,0.026592888765864905
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,65536,0.06468266910976835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,16384,128,0.005919111271699269
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3584,10240,0.027739556299315557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,32,2560,0.004567999806669024
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,512,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,768,64,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,8192,65536,0.13083555963304308
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,64,2560,0.006794666664467917
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,512,2560,0.0052284445199701525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,1536,0.012433778080675336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,2048,65536,0.06466044320000543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,8192,0.026904000176323786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,1024,16384,0.025252444876564875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,6144,0.011752888560295105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,8192,16384,0.02711822258101569
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,12288,12288,0.06534755229949951
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,128,0.009645333720578088
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,2560,7168,0.014725334114498563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,32,0.005271111097600725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,5120,0.03394844465785556
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,2560,1536,0.014448000325096978
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,512,16384,0.011740444435013665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,3584,1536,0.0069431112044387394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,128,2048,0.00630666646692488
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,7168,65536,0.12458577420976426
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,2560,0.007334222396214803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,1024,0.014120888378885059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,512,768,0.0028151110228565005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,4096,7168,0.019539554913838703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,768,0.02134399943881565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,8192,0.20859644148084852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,3072,0.0076453329788313965
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,8192,0.3395564556121826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,2560,0.03362400001949734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,512,0.0052391112678580815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,2048,128,0.0021742222209771476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,3072,7168,0.013646221823162503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,768,0.023414222730530634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,12288,4096,0.015807110402319167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,3072,8192,0.012808000048001608
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,5120,0.02803200152185228
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,3072,0.029380444023344252
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,2560,6144,0.009352000223265754
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,12288,0.05435644586881002
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,65536,0.025046222739749487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,128,0.006606222026877933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,65536,32,0.0041946665280395085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,16384,0.034380445877710976
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,10240,5120,0.019469334019554984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3584,512,0.008647999829716152
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,64,0.0021671110557185277
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,12288,7168,0.052912887599733144
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,12288,1024,0.007643555601437886
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,10240,64,0.019091554813914828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,10240,2048,0.00906488878859414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,8192,6144,0.018775999546051025
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,4096,6144,0.03357333276006911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,1536,2560,0.01831466621822781
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,3072,0.02083822256988949
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,2560,4096,0.023014222582181294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,6144,256,0.005903999838564131
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,16384,6144,0.03992799917856852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2560,16384,0.023427555958429973
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,64,6144,0.004906666775544484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,65536,256,0.013419555293189155
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,2048,8192,0.009673777553770278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,8192,2560,0.01000711073478063
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,2048,32,0.0021119999388853707
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,16384,0.19864000214470756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,12288,256,0.006234666539563074
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,10240,0.03074044320318434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,12288,0.12618045012156168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,64,1024,0.0064791109826829694
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,8192,1024,0.024892444411913555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,256,6144,0.030983110268910725
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,8192,768,0.004212444441186057
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,8192,16384,0.05548355645603604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,2048,0.031501332918802895
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,1536,0.006487111250559489
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,3584,0.017655111021465726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,768,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,6144,4096,0.02516888909869724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,12288,1024,0.0086666668454806
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,64,12288,0.014661333627170987
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,12288,4096,0.05253511004977756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2560,256,0.005872000008821487
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,2560,0.02084622283776601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,512,1024,0.013064888616402945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,256,6144,0.018194667167133756
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,2560,1024,0.0038488888078265717
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,256,32,0.006576889091067844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,128,3072,0.007255111303594377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,10240,1024,0.01036622209681405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,512,0.002516444358560774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,128,256,0.005984000033802456
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,128,0.015104888214005364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,7168,32,0.005912000106440649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,8192,0.014700444208251106
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,16384,32,0.0028577778074476454
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,2560,768,0.017847999930381775
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,7168,0.02441155579355028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,4096,2048,0.007167110840479533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,5120,65536,0.07477600044674344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,16384,1536,0.0285502208603753
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,5120,0.031508445739746094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,3072,256,0.0054551110499435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,1024,0.006606222026877933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2560,6144,0.011039999624093374
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,16384,2560,0.0165955556763543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,65536,0.06360089116626315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,7168,0.02004888819323646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,1024,10240,0.12735911210378012
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,5120,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,12288,12288,0.0358533329433865
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,3072,16384,0.019928000039524503
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,512,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,1024,0.028320001231299505
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,2048,32,0.005572444448868434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,768,0.0028444443725877335
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,3072,10240,0.0398613346947564
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,512,2560,0.004931555440028508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,5120,10240,0.015416888727082146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,6144,5120,0.014551111393504672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,4096,5120,0.2245617707570394
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,256,0.011376000112957425
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,12288,0.07870666848288642
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,3072,1024,0.0032177778581778207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1536,512,0.002514666567246119
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,3584,12288,0.017670222454600863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,2560,1536,0.006490666833188798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,64,128,0.0018337778747081757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,6144,6144,0.027021333575248718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,8192,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3584,65536,0.11087644762463039
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,2048,0.006584888945023219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,3072,3072,0.007890666524569193
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,128,0.0199297782447603
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,2560,0.007490666376219855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,2560,0.09519288937250774
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,16384,16384,0.042618665430280894
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,16384,10240,0.05561777618196276
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,16384,0.030466665824254353
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,32,768,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,512,256,0.0023146667000320223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,65536,256,0.007691555553012424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,512,12288,0.013078221844302284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,3584,6144,0.009314667019579146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,16384,6144,0.04182933436499702
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,512,0.011744000017642975
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,5120,0.023175999522209167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,65536,0.17957954936557344
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,768,0.01481244464715322
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,768,10240,0.012064889073371887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,512,1536,0.003506666670242945
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,16384,2048,0.012998222476906247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,4096,3072,0.0074879998962084455
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,4096,1024,0.016186666157510545
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,32,12288,0.005274666680230035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,8192,10240,0.05009955498907301
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,10240,2048,0.009863110880057016
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,3072,4096,0.14105155732896593
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,3072,32,0.001793777777089013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,8192,3584,0.011506666739781698
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,4096,0.01240355521440506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,3072,0.01146755533085929
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,1024,0.044828444719314575
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,128,32,0.005587555468082428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,3072,0.027035555905765955
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,5120,12288,0.036764442920684814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,12288,3072,0.01665599975321028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,32,0.005424889011515512
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,2048,0.01479200025399526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,5120,3072,0.023001778456899855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,6144,0.07230933507283528
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,1024,256,0.014885332849290637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,1536,0.01407911048995124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1536,256,0.006582222051090664
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,8192,6144,0.014139556222491793
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,256,3072,0.01566488875283135
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,3584,64,0.005612444546487596
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,16384,16384,0.9054604636298286
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,128,2048,0.017482666505707633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,16384,65536,0.22153511312272814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1536,768,0.024599111742443506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,32,0.0052533331844541765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,768,512,0.0028239999794297745
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,3584,0.007033777733643849
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,8192,256,0.00815644446346495
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,10240,5120,0.02499733368555705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,12288,2048,0.010732444624106089
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,512,1024,0.013162666724787818
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,7168,3584,0.13769955105251735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,1536,65536,0.5035937627156576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,10240,1024,0.00996800015370051
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,12288,8192,0.02404444416364034
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,12288,1024,0.037626667155159846
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,768,128,0.01071111112833023
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,7168,6144,0.015426667200194465
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,1024,0.01648355523745219
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,12288,0.03904622130923801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,32,0.0058808889653947614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,3072,2048,0.006890666567616993
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,512,3072,0.007263999846246507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,512,0.015153777268197803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,5120,0.011344888971911537
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,3072,0.007062222394678328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,8192,0.014879999889267815
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,2048,0.01828888886504703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,10240,0.011306666665607028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,1536,4096,0.008678221868144141
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,4096,128,0.0028159999185138275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,128,7168,0.00519466648499171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,2560,8192,0.020041777027977836
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,10240,0.012151111331250934
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,4096,512,0.005943111247486538
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,7168,0.2678915659586589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,12288,512,0.0069475554757648045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,12288,0.2385324372185601
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,2048,0.02718488872051239
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,128,0.005930666708283954
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,8192,0.11253955629136826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,768,7168,0.025807999902301367
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2560,7168,0.010008888940016428
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,768,4096,0.01720622181892395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,128,2048,0.012381333443853589
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,10240,512,0.04182489050759209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,256,16384,0.016488000750541687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,256,256,0.004921777794758479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,65536,7168,0.08567555745442708
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,2560,0.004902222090297275
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,32,64,0.0018328889790508484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,65536,1024,0.028823110792371962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,10240,6144,0.03970222340689765
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,65536,0.007054222126801808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,8192,1536,0.0069351109365622205
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,4096,8192,0.025758221745491028
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,8192,256,0.051382223765055336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,65536,1024,0.017795556121402316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,2560,1536,0.01648888819747501
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,64,8192,0.005210666606823604
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,8192,0.018561777141359117
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,3072,768,0.005595555735958948
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,16384,0.27358489566379124
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,65536,1536,0.06943733162350126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,16384,0.11978667312198216
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,16384,1024,0.008986666798591614
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,64,0.0018435555199782054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,768,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,2560,12288,0.028888887829250757
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,2048,1536,0.011824000212881299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,6144,0.009005332986513773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,512,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,7168,8192,0.03159200151761373
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,10240,2048,0.009697777529557547
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1536,4096,0.02163644466135237
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,3584,3584,0.008294222255547842
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,16384,64,0.0024888888001441956
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,768,10240,0.01272266689274046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,16384,0.10839288764529759
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3584,3584,0.008296888735559251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,1024,12288,0.011378666592968835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,10240,3072,0.049885332584381104
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,2048,12288,0.017280000779363845
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,65536,0.136771559715271
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,512,65536,0.06377244657940336
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,65536,0.04927911029921638
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,8192,0.01812177730931176
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,12288,2048,0.010657777388890585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,2048,5120,0.010029333333174387
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,4096,128,0.014821334017647637
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3072,4096,0.013521778086821238
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,6144,3072,0.031934221585591636
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,12288,2560,0.012088889049159156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,3584,2048,0.0069617778062820435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,2048,0.028220444917678833
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,32,128,0.001867555495765474
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,1536,0.021783111823929682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,3584,7168,0.01961688862906562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,1536,32,0.0018568889548381169
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,64,128,0.006654222392373615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1536,2560,0.004910222358173794
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1024,8192,0.027334221535258826
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,65536,2048,0.07698311408360799
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,8192,12288,0.032080001301235624
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,64,2048,0.006918222126033571
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,4096,6144,0.01240799989965227
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,5120,0.0306640002462599
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,7168,6144,0.012378666963842181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,10240,0.01836266616980235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,32,768,0.006256000035338932
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,7168,0.044694221682018705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,10240,3584,0.03693866729736328
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,1024,2048,0.016145777371194627
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,64,1536,0.0035439998739295532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,32,16384,0.0052471111218134565
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,1536,12288,0.016506666938463848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,32,5120,0.009325332939624786
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3584,7168,0.013274666335847644
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,65536,2048,0.029440889755884807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,2560,0.08304444286558363
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,4096,0.00830844458606508
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,65536,768,0.04073777794837952
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,10240,10240,0.03503466645876566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,8192,16384,0.03935111231274075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,1024,6144,0.009331555830107795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,2048,3584,0.01707555519209968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,2560,0.01888799998495314
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,1024,2048,0.00666222224632899
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,2048,3072,0.0076248885856734375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3584,3072,0.0196079992585712
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,7168,65536,0.07410844167073567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,64,12288,0.005468444277842839
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,768,12288,0.010022222167915767
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,5120,1536,0.006568888823191325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,16384,64,0.0021093332519133887
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,12288,7168,0.033082667324278094
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,65536,64,0.010332444475756751
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2048,8192,0.009317333499590555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,8192,0.014752000570297241
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,768,16384,0.016527111331621807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,4096,32,0.0021884443445338146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,8192,65536,0.09948177470101251
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,2048,1024,0.015816888875431485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,10240,0.13728266292148167
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,12288,2048,0.013223999904261695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,2048,0.0064186664919058485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,16384,3.185576968722873
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,80,6144,2560,0.007333333293596904
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,8192,0.021969777014520433
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,8192,3584,0.015494222442309061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,8192,6144,0.037441776858435735
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,12288,7168,0.04649777876006233
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,65536,5120,2.293940438164605
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,5120,7168,0.03715555535422431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,1536,4096,0.010434666441546546
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,1024,2048,0.02126755482620663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,6144,16384,0.04012177719010247
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,256,8192,0.023035556077957153
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,3584,1536,0.006274666637182236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,256,5120,0.009519110951158736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,5120,0.006270222365856171
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,4096,0.005111999809741974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,16384,3584,0.03805689016977946
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,256,7168,0.0052355556852287715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,3072,3072,0.007316444483068254
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,512,4096,0.046421332491768726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,768,0.00305244450767835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,2048,0.015146666102939181
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,1024,768,0.007287110719415877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,16384,4096,0.44755999247233075
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,256,65536,0.018256000346607633
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,256,3072,0.007295110987292395
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,10240,8192,0.025096888343493145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,256,0.01648177703221639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,512,65536,0.06054400073157417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,256,3072,0.005564444594913059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,6144,6144,0.018769777483410306
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,1024,65536,0.06146666738722059
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1024,3072,0.009611555271678502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,1536,0.07632444302241008
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,4096,1024,0.003495111233658261
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2560,6144,0.011730666789743634
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,5120,3584,0.008640000389681922
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,5120,0.03501510951254103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,5120,1024,0.016671111186345417
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,4096,1024,0.015437333120240105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,512,128,0.002540444541308615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,4096,64,0.0021155555215146807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,7168,0.014440889159838358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,6144,3584,0.009377777576446533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,64,256,0.002499555548032125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,128,7168,0.019564444820086162
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,3072,16384,0.03513866662979126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,12288,12288,0.04202933443917168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,256,12288,0.013387555877367655
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,1536,0.009868444667922126
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,12288,7168,0.034493333763546415
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,3584,0.009392889009581672
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,2560,768,0.012392889294359418
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,1536,4096,0.007967111137178209
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,6144,3584,0.007999111380841997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3072,3584,0.00866844422287411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,4096,2048,0.024567999773555334
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,128,2048,0.004628444297446145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,256,6144,0.009336888790130615
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,2048,10240,0.01554666625128852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,6144,32,0.0028782222006056043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,5120,0.897280904981825
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,256,512,0.015117333994971381
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2,7168,2048,0.006641777853171031
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1536,10240,0.015953777564896476
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,1536,32,0.0018257778137922287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,10240,3072,0.012712000144852532
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,64,0.001840888833006223
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,3584,768,0.006256888724035687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,6144,0.018603555030292936
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,65536,5120,0.050916443268458046
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,8192,3584,0.01202311118443807
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,1536,10240,0.011776000261306763
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,2048,1024,0.01679644485314687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,768,0.0026640000028742682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2048,2560,0.007279111279381647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,1024,6144,0.03645688957638211
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,10240,2048,0.028730667299694482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,8192,3584,0.03180444571706984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,2048,12288,0.016869333055284288
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,16384,5120,0.049950222174326576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,3584,2560,0.0517119997077518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,12288,3072,0.030619555049472388
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,7168,16384,0.02539288832081689
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,128,4096,0.0038497779104444715
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,12288,0.01752088963985443
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,3072,4096,0.025102222959200542
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,12288,65536,0.1653075615564982
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,2048,4096,0.009341333475377824
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,1024,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,64,0.03124533428086175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,1536,7168,0.010021333065297868
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,12288,64,0.035622222556008235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4096,512,10240,0.03433155682351854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,6144,1024,0.01680355601840549
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,768,0.01441422270403968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,8192,0.03321866525544061
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,16384,1536,0.07063288821114434
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,7168,768,0.003966222206751506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,65536,7168,0.14111377133263484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,65536,128,0.019542222221692402
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,16384,0.018210666047202218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,6144,0.018522666560279
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,6144,512,0.007259555160999298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,1536,0.019152889649073284
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,2048,2560,0.007289778027269576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,3072,256,0.002291555619902081
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,10240,10240,0.04091555542416043
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,5120,1024,0.010040000081062317
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,1024,2048,0.017224889662530687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,512,1024,0.012445333103338877
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,5120,5120,0.015815110670195687
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,256,128,0.013192888763215808
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,6144,0.02834933333926731
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,1536,0.020967110991477966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,5120,256,0.002867555452717675
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,64,65536,0.007262222468852997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,384,3584,7168,0.018960888187090557
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,1536,3072,0.015526221858130561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,12288,768,0.005606222069925732
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,10240,0.24369690153333876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,7168,0.05664177735646566
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,7168,64,0.0021848889688650766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,4096,2560,0.023567999402681988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,256,7168,0.009656888743241629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,1024,3584,0.0052693333062860705
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,1024,1536,0.003896000070704354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,12288,10240,0.0686968896124098
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,16384,64,0.009002666506502364
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,768,3584,0.007688889073001013
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,5120,1536,0.06796711020999484
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,64,64,0.0020355555332369274
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,64,12288,0.0052800000541739995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,1024,2048,0.016787555482652452
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,7168,512,0.016516443755891588
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,5120,7168,0.028334223561816748
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,8192,7168,0.017668444249365065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,2048,0.006612444503439798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,768,2560,0.007616889145639207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,4096,0.008020444048775567
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,5120,5120,0.016982222596804302
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,65536,16384,1.592831081814236
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,64,4096,0.012030222349696688
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,10240,3584,0.01793866687350803
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,8192,0.08286133077409533
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,1536,1536,0.003984000119898054
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,65536,128,0.010524444282054901
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,128,7168,0.02609599961174859
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,3072,512,0.002532444480392668
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,2048,1024,0.00629688882165485
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,1024,0.00509688879052798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,512,10240,0.0329039990901947
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,2048,7168,0.014071111877759298
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,7168,1536,0.01834133267402649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,512,16384,0.009568889107969072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,768,1536,0.01700088878472646
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,65536,0.15804355674319798
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,2560,32,0.0018835556175973679
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,3072,2048,0.006353777729802662
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,16384,12288,0.31738577948676217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,2048,0.020612445142534044
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,512,64,0.00184888889392217
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,6144,256,0.015623999966515435
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,3584,32,0.006573333508438534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,5120,64,0.005616888817813661
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,128,5120,0.008958222137557136
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,10240,12288,0.0456026660071479
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,2048,0.004544888933499654
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,128,1024,0.005608888963858287
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,2560,0.008568888737095727
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,1536,1024,0.006232000059551663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2048,768,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,256,6144,0.018239999810854595
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,32,1024,0.0028080000645584534
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,16384,512,0.10753244161605835
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,3072,3072,0.01962222158908844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,1024,128,0.0024951110697454875
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,16384,0.12230311499701606
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,2560,0.007315555380450354
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,10240,0.024704888463020325
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,1024,5120,0.009015999734401703
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,256,256,0.005277333160241445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,2048,32,0.0018479999982648427
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,16384,1024,0.008450667063395182
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,512,7168,0.011400000088744693
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,3584,0.021382222572962444
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,1536,1536,0.006598222172922558
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,3072,16384,0.04922755559285482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,8192,128,0.018206222189797294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,10240,16384,0.05654488669501411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,512,0.031657778554492526
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,12288,0.11619910928938125
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,4,64,16384,0.005530666559934616
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,65536,7168,0.2716017829047309
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,3072,5120,0.010696888797812991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,7168,1536,0.020592000749376085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,256,0.006264888991912206
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,5120,2560,0.023079999619060095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,768,256,0.002491555487116178
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,256,65536,0.09958933459387885
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,8192,0.015783111254374187
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,8192,1536,0.0069511110583941145
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,2560,0.011988444460762871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,128,65536,0.07949866851170857
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,3584,256,0.005570666657553778
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,1024,0.025432000557581585
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,65536,1536,0.1109608875380622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,8192,4096,0.06362666686375935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,512,512,0.002520888836847411
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,3584,0.16371377309163412
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,3584,0.055007110039393105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,1536,10240,0.008769777913888296
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,12288,2048,0.011042666931947073
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,128,0.0019511110666725372
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,256,6144,0.010279999838935005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,16,1536,3072,0.005573333137565189
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,256,512,65536,0.06511466370688544
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,3072,0.024687111377716064
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,64,8192,0.011716444459226398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,5120,7168,0.03737333416938782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,10240,1024,0.007263999846246507
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,768,2560,0.015171556009186639
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,256,1024,0.012426666915416718
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,4096,10240,0.01792622274822659
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,5120,65536,0.06950133376651339
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,3072,7168,0.01070755554570092
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,3072,1024,0.005967999911970562
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,768,65536,0.026141333911154006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,7168,768,0.035967999034457736
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,768,7168,0.019518221418062847
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,16384,0.0353999998834398
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,4096,7168,0.05691111087799072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,128,8192,0.04632533258861966
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,4096,2560,0.018210666047202218
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8192,256,7168,0.06146044201321072
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,65536,3584,0.10389866431554158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,256,65536,0.08069066868888007
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,3072,0.034953776333067156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,8192,12288,0.038416001531812884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,64,5120,0.00666133314371109
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,10240,12288,0.035233779086007014
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,3584,128,0.014456000592973499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,7168,64,0.0018613334331247543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,128,128,0.005219555563396878
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,6144,128,0.014459555347760519
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,4096,4096,0.012106666962305704
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,80,768,4096,0.01648266613483429
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,8192,512,0.007461333440409766
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,6144,1536,0.006636444479227066
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,256,10240,0.01240711079703437
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,768,6144,0.01014311114947001
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,192,1536,2560,0.0158933334880405
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2,3584,4096,0.02645155621899499
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,7168,8192,0.03255911005867852
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2,128,10240,0.0052177777720822235
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,3072,8192,0.03875199953715006
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,10240,16384,0.12784622775183782
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,8192,2560,0.030992888742023047
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,160,256,5120,0.00665866666369968
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,512,1024,0.007538666327794393
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,2560,64,0.005621333503060871
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,512,8192,0.021459556288189355
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,256,4096,0.01479822231663598
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,5120,2560,0.007641777396202087
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,65536,6144,0.12263377507527669
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,2560,1536,0.007977777885066139
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,128,0.03553244471549988
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,512,65536,1.042968855963813
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,6144,5120,0.024335111180941265
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,128,512,0.005222222043408289
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,12288,32,0.007954667011896769
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,768,128,0.005245333330498801
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,8192,3072,0.011095999843544431
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,5120,8192,0.048807111051347524
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,768,128,0.004946666873163647
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,256,6144,4096,0.02581866747803158
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,3072,12288,0.017291555802027386
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,768,512,0.0025066667132907445
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,1024,0.10078133477105035
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,7168,2560,0.007617777420414819
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,3584,2048,0.007706666986147563
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,2048,5120,0.010625777973069085
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,512,8192,0.021151999632517498
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,2560,4096,0.014811555544535318
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,65536,3072,2.1979278988308377
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,64,128,1024,0.0029502221279674103
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,128,2048,0.010696888797812991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,65536,3584,0.04283644424544441
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,2560,1024,0.005910222315125995
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,3584,0.1751342217127482
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2,10240,3584,0.013481777575280933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,32,2048,1536,0.005567111074924469
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,768,7168,0.018590221802393597
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,8,7168,256,0.005397333453098933
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,1536,256,0.006285333385070165
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,2560,12288,0.040039112170537315
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,6144,65536,0.134007109536065
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,3584,256,0.003866666720973121
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,1024,3072,0.00869777798652649
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,128,2048,0.012430222498046027
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,16384,0.053242666853798755
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1,256,3584,0.014789332946141561
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,512,1536,5120,0.010288000106811523
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,32,5120,0.009709333380063375
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,512,2560,0.006950221955776215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,2560,65536,0.055359999338785805
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,16,5120,8192,0.014101333088344999
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,768,7168,0.009769777456919352
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,128,12288,16384,0.04641777939266629
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,1,8192,6144,0.014332445131407844
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,512,256,16384,0.022520000735918682
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,768,128,0.00489066665371259
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,16,8192,10240,0.03498488995763991
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1,3584,65536,0.07615111271540324
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,1024,8192,0.013095999757448832
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,80,256,512,0.005632888939645555
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2048,256,0.005640000104904175
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,1024,7168,0.023603555228975084
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,64,768,0.0025519999778932994
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,16,4096,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,128,3584,0.005234666582610872
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,512,4096,0.03166133496496413
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,5120,1536,0.008273777862389883
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,2048,7168,256,0.009646221995353699
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,512,1024,4096,0.012058667010731168
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,32,256,0.0021457777669032416
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,64,7168,1024,0.017843556072976854
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,256,65536,0.061807109249962695
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,128,1024,0.0059057776298787855
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,1536,128,0.01478666729397244
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,256,0.005209777918126848
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,6144,0.07184800174501207
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,2048,128,3072,0.011041777829329172
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,128,128,768,0.013076444466908773
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,2560,1536,0.0069004446268081665
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,2048,0.03963111175431146
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,48,2560,16384,0.020710221595234342
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,32,10240,6144,0.031719999180899724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,128,128,0.0018044445249769424
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4,2560,3584,0.007606222397751278
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,5120,768,0.003858666867017746
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,1536,7168,0.01240355521440506
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,2560,4096,0.00794400026400884
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,3072,768,0.017162667380438913
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,7168,4096,0.0454382234149509
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,96,256,12288,0.006616888774765863
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,16384,10240,0.028780443800820246
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,256,2560,0.013960889644092984
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,2048,256,65536,0.18328711721632215
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,256,256,2560,0.004881777697139316
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,48,512,6144,0.01737511157989502
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,80,65536,8192,0.17677422364552817
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,2048,0.03360177742110358
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,6144,1024,0.006252444452709622
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,6144,256,0.015819556183285184
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,12288,6144,0.053007112609015576
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4,512,4096,0.017498667041460674
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,512,1024,0.005886222339338726
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,768,2048,0.014521777629852295
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,48,1024,12288,0.015108444624476962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,64,2560,0.004860444615284602
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,32,2560,7168,0.013064000341627332
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,32,3072,12288,0.017056889004177518
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,768,10240,768,0.01370044466522005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1024,512,65536,0.13148710462782118
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,192,512,4096,0.00832711077398724
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,384,5120,4096,0.016359999775886536
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,96,5120,1536,0.006585777633719974
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,64,128,768,0.0063457778758472866
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,64,768,768,0.005590222362014983
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,8,6144,7168,0.035939554373423256
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,768,12288,768,0.011647111011876
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8192,7168,5120,0.3971377743615045
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,48,16384,512,0.004523555437723796
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,384,3072,512,0.005928888916969299
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,160,65536,2048,0.037504888243145414
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,8,10240,1024,0.005881777654091517
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,96,5120,128,0.00573511090543535
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4,4096,65536,0.06778755452897814
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,1,1024,65536,0.05739111370510525
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,192,64,6144,0.005264889034960005
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,1024,512,512,0.013447110851605734
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,3584,0.1054311129781935
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8,4096,6144,16384,0.38623732990688747
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,160,3072,2048,0.006931555353932911
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,1024,2560,7168,0.026965333355797663
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,768,8192,65536,0.5534719891018337
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,192,12288,128,0.006569777925809224
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,160,12288,128,0.015787555111779105
SGLang,0.5.9,NVIDIA B200,gemm,sglang,float16,128,65536,1536,0.04071733355522156
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,96,7168,5120,0.026611554953787062
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,8,7168,1536,0.006265777680608962
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,768,0.042232000165515475
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,3072,0.04691911074850294
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,384,7168,7168,0.038732445902294584
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,128,65536,2560,0.0273671117093828
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,256,65536,8192,0.07427555322647095
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,768,2560,7168,0.03568444318241543
SGLang,0.5.9,NVIDIA B200,gemm,sglang,fp8_block,512,1024,5120,0.02240088913175795
SGLang,0.5.9,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,3072,0.3354453245798747
