framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,2560,0.0064293332397937775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,5120,0.03827555643187629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,2048,0.009029332962301042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,512,0.00564533347884814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,3072,0.015271999769740634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,10240,0.014039110806253223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,128,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,6144,0.013591999808947245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,512,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,10240,0.0500311123000251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,256,0.003546666767862108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,512,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,3072,0.2340328958299425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,2048,0.0070266665683852295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,768,0.006160888820886612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,3072,0.011805333197116852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,1024,0.005594666633341048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,3584,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,2048,0.04172177778349983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,1536,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,2560,0.010867555936177572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,8192,0.010374222364690568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,2048,0.007215110792054071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,10240,0.03432177835040622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,6144,0.023223110371165808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,3584,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,2560,0.007326222128338284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,5120,0.01793333391348521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,256,0.006262222097979651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,16384,0.057690666781531445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,1024,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,65536,3.9170345730251737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,5120,0.0148026661740409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,16384,0.3896595637003581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,7168,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,3584,0.010041777458455827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,10240,0.04275822308328417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,768,0.005571555760171678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,2560,0.015498666299713982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,128,0.01484266585773892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,128,0.006324444380071428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,10240,0.04177511069509718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,16384,0.01573244399494595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,8192,0.9675359725952148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,6144,0.0746284458372328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,2048,0.00885333369175593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,3072,0.007342221836249034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,64,0.008358221915033128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,1024,0.01035911093155543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,1024,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,512,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,12288,0.013092444174819522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,256,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,10240,0.016955556141005624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,128,0.002850666642189026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,1024,0.04239555531077915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,1024,0.014191110928853353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,512,0.01796800063716041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,64,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,2048,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,4096,0.008330666356616551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,2048,0.09171466694937812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,16384,0.02370311154259576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,7168,0.009716444545321995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,256,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,4096,0.018636445204416912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,4096,0.008580444587601556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,768,0.0039022221333450745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,2048,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,10240,0.017163554827372234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,12288,0.020019556085268658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,2560,0.032136887311935425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,1536,0.027128888501061335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,128,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,64,0.007367110914654202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,3072,0.01071111112833023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,1536,0.006743999818960826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,3584,0.0158915552828047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,1536,0.1663866705364651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,2048,0.0210826661851671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,16384,0.02646666599644555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,7168,0.025158221522967022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,128,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,2560,0.015915556086434256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,1536,0.016927111479971144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,512,0.007991111112965478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,256,0.021686222818162706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,16384,0.02388266722361247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,512,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,256,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,64,0.008325333396593729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,3072,0.01290933373901579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,65536,0.05672711133956909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,256,0.0032417778339650896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,65536,0.06820799907048543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,64,0.011866666376590729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,8192,0.01514311134815216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,1024,0.014860444598727755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,32,0.02444444431198968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,4096,0.005555555638339784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,2560,0.0318577786286672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,512,0.005223999834722943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,512,0.004598222259018156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,5120,0.01590755581855774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,2560,0.009276444713274637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,3072,0.06582488616307576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,2560,0.07815911372502644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,12288,0.032066666417651706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,1536,0.004951111144489712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,5120,0.020987555384635925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,256,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,128,0.00977422214216656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,5120,0.06745688782797919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,3584,0.018329777651362948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,5120,0.05232977867126465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,2048,0.012108444339699216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,64,0.0031999999450312722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,16384,0.09058577484554714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,3584,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,5120,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,256,0.014450665977266101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,1024,0.014165333575672574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,3072,0.02628622286849552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,10240,0.017698667115635343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,5120,0.020779555042584736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,512,0.00830488900343577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,1024,0.0038248888320393036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,2560,0.01664177742269304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,8192,0.07148977782991198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,1024,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,16384,0.015492444237073263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,1536,0.02772177755832672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,128,0.012505778008037143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,1536,0.01514933341079288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,8192,0.009034666750166152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,256,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,5120,0.010779555473062726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,3072,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,65536,0.09312266773647732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,65536,0.05949866771697998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,32,0.0036595554815398324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,7168,0.012056888805495368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,64,0.006714666469229593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,1024,0.03171111146608988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,10240,0.05485688977771335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,128,0.014831110835075378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,256,0.0032319999817344877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,2048,0.007310222420427535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,12288,0.052870220608181424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,3072,0.03314044409328037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,512,0.003882666842805015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,5120,0.015839111473825242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,768,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,1024,0.020967110991477966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,1536,0.01977688901954227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,16384,0.01584533353646596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,6144,0.15165689256456164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,2560,0.007003555695215861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,768,0.005969777703285217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,3072,0.05626489056481255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,8192,0.256660434934828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,5120,0.016833777228991192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,7168,0.030593779351976182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,32,0.0032417778339650896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,16384,0.01552177800072564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,3584,0.009031111167536842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,5120,0.009452444811662039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,4096,0.03893866803910997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,1536,0.015215999550289579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,32,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,10240,0.09113688601387872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,64,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,3584,0.08180889156129625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,32,0.013968888256284924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,6144,0.012796444197495779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,128,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,3072,0.015818667080667283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,2560,0.018944889307022095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,12288,0.023456000619464453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,1024,0.01604622271325853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,2560,0.02115999990039402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,10240,0.03333244389957852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,8192,0.021714667479197185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,2048,0.012785777449607849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,65536,0.15391466352674696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,32,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,10240,0.034244444635179304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,5120,0.040896889236238264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,16384,0.026218665970696345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,3584,0.01680355601840549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,7168,0.02494488822089301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,128,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,256,0.011757333245542316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,2560,0.015246222416559855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,2048,0.015656888484954834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,768,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,2560,0.026404443714353774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,128,0.005232000102599462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,2048,0.00793599999613232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,4096,0.00792622235086229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,65536,0.0796017779244317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,3584,0.009200889203283522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,12288,0.06648177570766874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,5120,0.015374221735530429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,2048,0.029765334394243028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,65536,0.06086044179068672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,2560,0.012201777762836881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,6144,0.05875555674235026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,4096,0.012085333466529846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,8192,0.027791112661361694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,1536,0.0063279999627007385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,512,0.016884444488419425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,12288,0.013441777891582914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,32,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,512,0.003903999924659729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,10240,0.10232888989978367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,7168,0.011404444773991903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,4096,0.019309333629078336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,7168,0.017284444636768766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,128,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,512,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,65536,0.06332621971766154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,256,0.012947555217478009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,256,0.009315555294354757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,65536,0.05772711171044243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,512,0.012467555701732635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,16384,0.00868800034125646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,7168,0.009731555150614845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,64,0.006955555743641323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,32,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,1024,0.018573333819707234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,512,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,3072,0.014826666977670459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,3072,0.020081778367360432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,10240,0.012042666474978128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,768,0.008632889224423302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,3584,0.020581333173645865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,10240,0.022690667046440974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,7168,0.01757777730623881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,2048,0.0069466663731469055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,7168,0.01500888831085629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,65536,0.07423466444015503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,4096,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,256,0.005554666535721884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,256,0.00943911075592041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,10240,0.017642666896184284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,2048,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,768,0.01759022143152025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,5120,0.01536266671286689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,512,0.11759732829199897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,1536,0.004955555415815777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,10240,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,3584,0.016015110744370353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,2048,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,2048,0.006450666735569636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,64,0.014884443746672736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,12288,0.03742400142881606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,512,0.01404711107412974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,256,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,16384,0.08095110787285699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,8192,0.014120888378885059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,16384,0.08595199717415704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,256,0.03547733359866672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,128,0.003165333428316646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,10240,0.018182223041852314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,16384,0.04775466521581014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,768,0.0035013332962989807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,2048,0.02094399929046631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,64,0.003859555555714501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,768,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,6144,0.01723288827472263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,16384,0.07233777973386976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,2560,0.009372444616423713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,2560,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,4096,0.019704888264338177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,7168,0.009672889278994666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,6144,0.01428533262676663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,6144,0.02664622167746226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,2560,0.011924444801277585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,8192,0.02068711154990726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,3072,0.022252443763944838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,2048,0.028218666712443035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,4096,0.00794577764140235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,16384,0.022946667340066697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,512,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,768,0.009209777745935652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,4096,0.007676444119877285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,256,0.01866755551762051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,2048,0.01037599974208408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,1024,0.0038808890514903595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,1536,0.052854223383797534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,10240,0.13181155257754856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,2560,0.014847111370828418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,3072,0.016833777228991192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,512,0.01352266636159685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,2048,0.007310222420427535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,2560,0.039911998642815485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,512,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,1024,0.01276711126168569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,8192,0.18095378081003824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,65536,0.14690489239162868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,2560,0.013352889153692456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,3584,0.013181333740552267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,5120,0.008391111261314815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,2048,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,3072,0.08338933520846897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,768,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,512,0.35859200689527726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,8192,0.012464000119103326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,3072,0.11516622702280681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,512,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,3072,0.006127999888526068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,512,0.017208000024159748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,6144,0.02362577782736884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,32,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,65536,0.1095599995719062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,16384,0.019631110959582858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,12288,0.027232888672086928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,4096,0.011379555695586734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,16384,0.048037333620919116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,512,0.0038515557017591265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,12288,0.030138668086793687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,2560,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,6144,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,3072,0.005630222045713001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,6144,0.0564862224790785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,5120,0.2133457793129815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,1024,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,64,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,16384,0.02091377807988061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,1024,0.017443555924627516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,3584,0.013915555344687568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,16384,0.038839111725489296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,1536,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,65536,0.12180799908108181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,12288,0.031199110878838435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,65536,0.146342224544949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,1536,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,1024,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,1536,0.015519110692871941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,2560,0.01867733399073283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,3072,0.08730844656626384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,3584,0.014499555031458536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,1536,0.01587466730011834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,8192,0.011058666639857821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,4096,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,65536,0.04877955714861552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,128,0.015478221906556023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,12288,0.07830133040746053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,32,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,768,0.0063582224150498705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,10240,0.015799999237060547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,8192,0.043997334109412305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,512,0.016807110773192514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,512,0.018566222654448617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,10240,0.12991555531819662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,2048,0.006586666736337874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,8192,0.03348088926739163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,256,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,10240,0.03107644451989068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,1536,0.5890293121337891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,3584,0.10767910877863567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,65536,0.7368897861904569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,2048,0.05979910824033949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,7168,0.010779555473062726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,512,0.016198222835858662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,512,0.029784000582165186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,1024,0.01514400045077006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,6144,0.021398221453030903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,2560,0.0660053359137641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,65536,0.17578489250606963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,1024,0.004554666578769684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,5120,0.01607466737429301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,2048,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,256,0.010320000350475311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,32,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,1536,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,7168,0.007680888805124495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,65536,0.08305866850747003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,12288,0.00830488900343577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,3072,0.007595555649863348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,2560,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,12288,0.06605688730875652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,3072,0.03695911169052124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,128,0.014514666464593677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,3072,0.007352888584136963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,64,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,1024,0.013456888496875763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,512,0.005234666582610872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,3072,0.00903911143541336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,3072,0.009318222602208456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,65536,0.2645484341515435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,16384,0.03125244379043579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,10240,0.03658399979273478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,2048,0.01756444407833947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,8192,0.06650133265389337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,256,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,128,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,6144,0.028777778148651123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,10240,0.01575644479857551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,1024,0.015855110353893705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,128,0.031697776582505964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,4096,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,3584,0.008065777520338694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,128,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,1024,0.01885511146651374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,10240,0.027477333943049114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,256,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,256,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,256,0.00286666676402092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,10240,0.047573334640926786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,4096,0.015989333391189575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,2048,0.024528889192475214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,1024,0.01789422167672051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,3072,0.008351999852392409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,64,0.002789333255754577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,5120,0.021720000439220007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,16384,0.016935999194780987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,768,0.016855110724767048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,768,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,64,0.003625777860482534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,3072,0.013008000122176277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,6144,0.10999021927515666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,8192,0.18108977211846244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,5120,0.01314222233162986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,65536,0.06410133176379733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,10240,0.014481777946154276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,512,0.0035235554807715942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,1536,0.022711111439598933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,8192,0.05888533592224121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,512,0.06847200128767225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,768,0.005985777825117111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,8192,0.02529777752028571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,65536,0.10963555839326646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,1024,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,1024,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,1024,0.1563297775056627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,3072,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,16384,0.0891324414147271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,64,0.0028364445186323593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,5120,0.053156445423762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,3584,0.01295111080010732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,10240,0.01166488892502255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,1536,0.011600889265537262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,768,0.0035093331502543557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,16384,0.0336168905099233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,8192,0.03316977620124817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,2048,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,10240,0.023510222633679707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,1024,0.015121777852376303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,5120,0.08189866940180461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,65536,0.063701331615448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,65536,0.1550275617175632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,1536,0.010737777584128909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,128,0.00527555536892679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,7168,0.013394667042626275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,10240,0.012541333006487953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,8192,0.010681777364677854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,3584,0.017774222625626456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,512,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,7168,0.038741333617104426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,8192,0.027170666389995154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,128,0.003913777983850903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,16384,0.015807110402319167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,2560,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,8192,0.009659555223253038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,10240,0.01961955593691932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,768,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,256,0.0053013331360287136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,6144,0.021127111381954614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,1024,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,5120,0.12833421760135227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,2560,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,65536,0.021568000316619873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,10240,0.03421333432197571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,4096,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,10240,0.011363555159833698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,256,0.010062221851613786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,2560,0.015186667442321777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,65536,0.06533333328035143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,1024,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,1536,0.008629333641793992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,512,0.016919111212094624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,512,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,256,0.017190222938855488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,768,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,128,0.002899555489420891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,5120,0.012482666307025485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,256,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,5120,0.010079111489984725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,12288,0.03429866830507914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,65536,0.07408977879418267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,1024,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,12288,0.03774844606717428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,3072,0.01089866707722346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,10240,0.09358844492170547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,8192,0.0147733340660731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,256,0.005225777626037598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,2560,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,3072,0.010013333625263637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,64,0.0064035554726918536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,65536,0.05897066990534464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,65536,0.032368001010682844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,65536,0.08354489008585612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,4096,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,12288,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,256,0.005249777601824866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,256,0.005863111052248213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,3072,0.03414577907986111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,2048,0.024837333295080397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,3072,0.008512000242869059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,7168,0.019308444526460435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,768,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,65536,0.17512178421020508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,768,0.003858666867017746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,1536,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,1024,0.012808889150619507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,10240,0.02548533346917894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,3584,0.01997422178586324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,10240,0.012339555554919772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,512,0.006803555621041193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,2560,0.0295448899269104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,6144,0.09694221946928237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,7168,0.03626666797531976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,65536,0.07267289029227363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,2048,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,128,0.002999111182159848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,2048,0.023880889018376667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,64,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,8192,0.01853155593077342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,5120,0.012411555482281579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,32,0.003921777837806278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,768,0.0038533334930737815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,65536,3.6950976053873696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,3584,0.01651822196112739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,65536,0.029097778929604426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,3584,0.019609777463806998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,3584,0.025563556287023757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,7168,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,65536,0.10425600078370835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,128,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,4096,0.008272888759771982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,2048,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,128,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,4096,0.04464711083306206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,6144,0.018236445056067575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,3072,0.010231110784742568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,2048,0.01309333327743742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,4096,0.028661333852344092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,32,0.0025395556456512874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,12288,0.02699377801683214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,32,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,7168,0.27912622027926975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,3072,0.024866667058732774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,1024,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,512,0.012856889102194043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,16384,0.0255724440018336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,512,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,16384,0.05470666620466444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,5120,0.024120888776249353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,3072,0.03700266612900628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,7168,0.029389331738154095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,7168,0.023050665855407715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,7168,0.10909777879714966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,12288,0.01293511109219657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,6144,0.01688799924320645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,12288,0.06363377968470256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,7168,0.3472302224900987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,8192,0.04136088821623061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,6144,0.012823111481136747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,1024,0.003530666646030214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,512,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,12288,0.012558221817016602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,12288,0.05680711401833428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,5120,0.05003199974695841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,2560,0.026103110777007207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,12288,0.04037066631846958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,8192,0.008834666675991481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,3584,0.007623111208279927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,4096,0.008310221963458592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,5120,0.0206631107462777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,5120,0.007977777885066139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,7168,0.03237688872549269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,3584,0.012802666260136498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,768,0.003867555409669876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,2560,0.028816001282797918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,4096,0.0233297778500451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,4096,0.009755555954244401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,6144,0.06735555330912273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,8192,0.13872533374362522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,3072,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,7168,0.017878222796652053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,2560,0.017921777235137094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,2560,0.018160000443458557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,256,0.00527555536892679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,4096,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,65536,0.13004622194502088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,256,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,1024,0.03702400128046671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,65536,0.15708800156911215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,5120,0.02366133365366194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,2048,0.007992888490358988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,256,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,10240,0.011754666765530905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,64,0.006364444477690591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,6144,0.01795288920402527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,65536,0.05691288577185737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,512,0.015507555670208402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,64,0.004206222378545337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,65536,0.062231997648874916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,1536,0.012763555679056378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,7168,0.031376888354619346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,2048,0.011745777395036487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,1024,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,65536,0.06716710991329618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,2048,0.07810578081342909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,4096,0.018585777944988675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,16384,1.5368613137139215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,768,0.021535111798180476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,128,0.01143022212717268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,12288,0.02845066785812378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,5120,0.029194666279686823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,2560,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,512,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,512,0.005247999810510212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,768,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,2560,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,65536,0.19591377841101754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,3072,0.06066311068005032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,512,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,8192,0.01330488920211792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,7168,0.05910488631990221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,64,0.028773334291246202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,12288,0.027564444475703772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,5120,0.012151111331250934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,3072,0.024013333850436743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,10240,0.15792355272505018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,512,0.013404444687896304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,128,0.00314666661951277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,5120,0.008383110993438298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,2048,0.08460888597700332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,6144,0.01721244388156467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,256,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,4096,0.007826666865083907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,5120,0.009352888498041365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,7168,0.00795733349190818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,10240,0.026724444495307073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,3584,0.007895111209816402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,4096,0.01313777764638265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,65536,0.08988089031643337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,5120,0.23807999822828504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,1536,0.008168000313970778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,1024,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,7168,0.01593777702914344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,2560,0.014134221606784396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,6144,0.013479111095269522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,1536,0.028579556279712256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,768,0.004588444613748127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,64,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,7168,0.0635991096496582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,256,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,16384,0.020941333638297185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,16384,0.042912890513738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,8192,0.24236533376905653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,3584,0.014767110347747803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,512,0.005521777603361342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,7168,0.020301333732075162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,1024,0.0042053332759274375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,7168,0.061887111928727895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,5120,0.023736889163653057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,1024,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,64,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,65536,0.619832886589898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,768,0.027308444182078045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,10240,0.012512000070677863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,3072,0.010271111296282874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,3584,0.021077333225144282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,2560,0.012087999946541257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,12288,0.016912889149453905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,768,0.017301332619455125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,5120,0.00868888861603207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,128,0.0031911111954185697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,32,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,6144,0.01105955574247572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,64,0.002858666703104973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,1024,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,1024,0.007869333028793335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,3072,0.03185155656602647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,3072,0.058021333482530385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,10240,0.02994489007525974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,1536,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,4096,0.019159111711714003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,2048,0.012770666844315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,768,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,768,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,6144,0.012438221938080259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,12288,0.023649776975313824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,768,0.016207999653286405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,65536,0.06924266947640313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,1024,0.003920888735188378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,65536,0.8440524207221137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,8192,0.017267554998397827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,7168,0.022647110952271357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,256,0.00978488889005449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,1024,0.015503111812803479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,8192,0.02398577829202016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,3584,0.026576000783178542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,2048,0.011810666984981961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,64,0.003220444545149803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,3072,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,8192,0.030993776188956365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,256,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,8192,0.045886221859190196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,3072,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,6144,0.026478222674793665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,3584,0.019945777124828763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,6144,0.03256533212131924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,4096,0.014841778410805596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,512,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,3584,0.016140444411171805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,768,0.003942222230964237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,128,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,2048,0.011400000088744693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,256,0.009027555584907532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,3584,0.007624000310897827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,6144,0.009025777379671732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,2560,0.021212443709373474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,2048,0.01034755590889189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,5120,0.012587555580668978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,3072,0.00794400026400884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,128,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,64,0.00318755561278926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,12288,0.02769244379467434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,3072,0.027341332700517442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,1024,0.014104000396198697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,256,0.013368888861603208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,6144,0.012863111164834766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,8192,0.1524124410417345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,2560,0.010432889064153036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,768,0.029446221060223047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,1024,0.014979556202888489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,128,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,6144,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,1536,0.008922666311264038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,512,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,256,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,1536,0.009020444419648912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,512,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,2560,0.01756355497572157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,512,0.012771555946932899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,512,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,12288,0.02163733376397027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,4096,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,10240,0.012063999970753988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,5120,0.00794933322403166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,256,0.003527111063400904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,16384,0.02334755493534936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,1536,0.009604444106419882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,1024,0.0038284444146686127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,65536,6.628209855821397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,5120,0.045590221881866455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,2560,0.0069715554515520734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,8192,0.007298666569921706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,2048,0.06768266359965007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,64,0.010965333216720156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,32,0.0236826671494378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,10240,0.009189333352777693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,256,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,12288,0.12439200613233779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,3584,0.01611555616060893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,8192,0.13288622432284886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,4096,0.021546666820844013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,7168,0.03345155715942383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,7168,0.07732533084021674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,8192,0.010425777898894416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,1536,0.010199999643696679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,512,0.011771555576059552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,65536,0.024904888537194993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,12288,0.017831110292010836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,768,0.011448889142937131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,1536,0.018817777434984844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,6144,0.010840000377760993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,32,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,6144,0.013239111337396832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,512,0.007993777592976889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,1536,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,128,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,6144,0.01038755559258991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,64,0.0028328889360030494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,3072,0.11167466640472412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,4096,0.015092444088723926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,512,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,5120,0.009357333183288574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,2560,0.032300445768568255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,256,0.007956444389290279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,128,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,512,0.014116444521480136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,2048,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,8192,0.010050666828950247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,2048,0.0063440000845326324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,128,0.019139554765489366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,64,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,16384,0.02087466584311591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,7168,0.03218311071395874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,32,0.005534222142563925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,10240,0.01717866626050737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,3584,0.017845332622528076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,32,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,3584,0.013307555682129331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,8192,0.30003645684984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,3072,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,32,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,7168,0.011051555474599203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,5120,0.4445670975579156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,768,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,768,0.00703999979628457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,128,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,4096,0.01793866687350803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,1536,0.011428444749779172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,1536,0.007239110767841339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,2560,0.011724444727102915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,5120,0.012802666260136498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,5120,0.015834665960735746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,7168,0.020768000019921195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,64,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,2560,0.019897777173254225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,4096,0.007999111380841997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,1536,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,128,0.003495111233658261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,65536,0.0635440018441942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,2048,0.011236444115638733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,768,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,256,0.011426666544543372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,2560,0.009022222624884712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,65536,0.09860089090135361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,3072,0.0075342224703894714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,12288,0.02847822176085578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,512,0.009014222357008193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,5120,0.00867022242810991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,12288,0.02607822252644433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,2560,0.00812177773978975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,2560,0.056681778695848256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,16384,0.7209733327229818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,7168,0.009663999908500248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,2560,0.03263555632697211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,12288,0.028656889994939167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,8192,0.3081706629859077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,256,0.03331822156906128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,1024,0.006311111152172089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,128,0.00980000032318963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,1024,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,7168,0.01344622257683012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,32,0.0058782220714622075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,65536,0.3702497747209337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,1024,0.003980444537268745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,8192,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,5120,0.008720888859695857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,3584,0.014826666977670459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,2560,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,64,0.008664888640244802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,12288,0.04180711176660326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,2560,0.02964799933963352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,128,0.0031848889258172777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,32,0.0034933334423436057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,65536,0.07602844635645549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,3072,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,3072,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,256,0.01687733332316081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,8192,0.02551555633544922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,3584,0.013723555538389417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,7168,0.011377777490350934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,10240,0.013778666655222574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,65536,0.11680355336931016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,5120,0.011180444724029966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,1536,0.004560889055331548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,1536,0.013832000394662222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,12288,0.018963555494944256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,3072,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,768,0.012451555165979596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,16384,0.022866666316986084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,10240,0.022388445006476507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,2560,0.015798222687509324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,7168,0.018696000178654987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,7168,0.03580888774659898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,16384,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,10240,0.04475288920932346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,3072,0.028530667225519817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,7168,0.0732826656765408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,768,0.004192000048028098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,12288,0.22819643550448945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,4096,0.025575111309687298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,8192,0.03652088840802511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,8192,0.030035555362701416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,768,0.02880177895228068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,5120,0.022142221530278523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,768,0.007381333245171442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,512,0.013208000196350945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,128,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,2048,0.00620177760720253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,512,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,128,0.0057555557125144536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,12288,0.01794577803876665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,65536,0.08077866501278348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,16384,0.04901955525080363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,10240,0.0158924443854226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,8192,0.12131378385755752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,3072,0.020987555384635925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,6144,0.00666133314371109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,3072,0.02230311102337307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,7168,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,8192,9.836815728081598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,128,0.011711111499203576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,256,0.01520088811715444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,768,0.01716711123784383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,3584,0.0074826669361856245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,512,0.012695999609099494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,768,0.00350577798154619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,1536,0.019703111714786954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,512,0.003540444291300244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,2560,0.0058808889653947614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,3584,0.011545778148704104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,1024,0.05150399936570061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,3584,0.008418666819731394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,1536,0.023938667443063524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,5120,0.042019555966059365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,768,0.06851466496785481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,256,0.013601777454217276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,2048,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,3584,0.009707555174827576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,10240,0.011904888682895236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,2560,0.011737777955002256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,768,0.008759111166000366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,256,0.014466666513019137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,16384,0.07444000244140625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,65536,0.06285155481762357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,1024,0.013381332986884646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,256,0.0028906667398081887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,16384,0.021338666478792827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,1536,0.004599110947714912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,16384,0.021809778279728357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,5120,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,4096,0.008010666403505538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,16384,0.02776799930466546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,10240,0.04294044441647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,2048,0.01198755535814497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,5120,0.008647110727098253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,3584,0.019242667489581637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,10240,0.012440889245933957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,1536,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,512,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,7168,0.03010933266745673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,10240,0.011368888947698804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,10240,0.01623911162217458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,256,0.07018133004506429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,512,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,128,0.015818667080667283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,1536,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,768,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,7168,0.03210577699873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,65536,0.1337626642651028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,768,0.006978666616810693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,256,0.005230222311284807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,1536,0.005306666509972678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,128,0.01620533400111728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,4096,0.012070222033394707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,6144,0.010771555205186209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,65536,0.15119732750786677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,256,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,32,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,512,0.0035706667436493766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,256,0.01625155574745602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,1536,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,16384,0.052490668164359204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,1024,0.004173333446184794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,1024,0.006288888967699475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,3584,0.02052888936466641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,16384,0.06899822420544095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,2560,0.010013333625263637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,64,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,12288,0.012455110748608908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,1024,0.018927110566033255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,512,0.012809777425395118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,3072,0.0063528890411059065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,2048,0.008627555436558193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,768,0.016501333978441026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,2560,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,128,0.009011555876996782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,7168,0.07564266522725423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,768,0.015793777174419828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,6144,0.017640888690948486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,12288,0.031143112315071955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,7168,0.02520977788501316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,2048,0.009147555463843876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,6144,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,256,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,10240,0.09684355391396417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,1536,0.009406222237481011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,12288,0.09440622064802383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,6144,0.014688889185587565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,16384,0.16147022777133518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,5120,0.006384000182151794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,128,0.005412444472312927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,10240,0.20611022578345406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,768,0.01775911119249132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,256,0.005707555760939916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,256,0.01517599986659156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,5120,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,3584,0.018956444329685636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,6144,0.021550221575631037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,3584,0.006945778098371293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,3584,0.0076044441925154785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,768,0.0057297779454125305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,1024,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,2560,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,2048,0.013521778086821238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,7168,0.02748888896571265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,2048,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,128,0.0038515557017591265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,3072,0.007278222176763747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,3584,0.007962666451931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,64,0.003903111235962974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,64,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,16384,0.22252533170912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,768,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,12288,0.06421422296100192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,7168,0.03489777776930068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,128,0.013493333425786761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,3584,0.06673422124650744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,1024,0.028068443139394123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,16384,0.013680000272062091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,32,0.002502222235004107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,2048,0.013281777501106262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,512,0.003463111196955045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,512,0.013434666726324292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,5120,0.010098666780524785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,16384,0.016167999969588388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,12288,0.017719111508793302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,2560,0.01015466699997584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,768,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,3584,0.01369244439734353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,768,0.01909244391653273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,2560,0.016672000288963318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,1024,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,16384,0.02127911150455475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,32,0.006324444380071428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,3584,0.00778311077091429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,65536,0.2449644406636556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,10240,0.22272443771362305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,256,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,1024,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,256,0.014858666393491956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,1024,0.025349333882331848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,2048,0.02238133384121789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,256,0.006633777585294511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,6144,0.01128799964984258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,65536,0.07205955849753486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,32,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,16384,0.144813338915507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,65536,0.7350435786777072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,32,0.11772888236575657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,16384,0.014440000057220459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,3072,0.010315555665228102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,1536,0.008655110994974772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,768,0.010024888647927178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,1024,0.03438222077157762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,128,0.014956444501876831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,12288,0.027477333943049114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,2560,0.01552177800072564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,6144,0.028904000918070476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,128,0.014850666125615438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,768,0.016520889268981088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,1024,0.004126222183307012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,5120,0.012815111213260226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,4096,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,1024,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,6144,0.019428445233239066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,256,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,65536,0.1250213384628296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,768,0.004599110947714912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,6144,0.013942221800486246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,1536,0.01163822246922387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,2048,0.044976890087127686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,512,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,3072,0.017636444833543565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,1024,0.0041884444653987885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,1024,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,10240,0.04215822286076016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,4096,0.007919111185603673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,64,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,512,0.015127999915017022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,32,0.006319111006127463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,16384,0.05064710974693298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,2560,0.015535111228624979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,12288,0.04265333215395609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,65536,0.0808382232983907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,128,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,7168,0.009662222531106737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,3072,0.013012444807423485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,2048,0.014825777875052558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,4096,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,512,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,4096,0.012279111478063794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,1536,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,65536,0.17727200190226236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,5120,0.008901333643330468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,65536,0.1293626626332601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,64,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,1024,0.016193777322769165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,5120,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,2560,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,3072,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,10240,0.01752711170249515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,1024,0.007355555891990662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,4096,0.007630222373538547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,5120,2.870873769124349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,3584,0.01611199975013733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,65536,0.07488889164394803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,5120,0.008026666939258575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,512,0.01479288935661316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,7168,0.022337777747048274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,1024,0.005516444643338521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,12288,0.07608799801932441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,32,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,5120,0.12995644410451254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,12288,0.027851555082533095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,16384,0.34372178713480633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,1536,0.01181777815024058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,3072,0.01702666613790724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,5120,0.09221244520611233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,4096,0.011720889144473605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,12288,0.049457778533299766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,2048,0.01389955480893453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,7168,0.010223111344708337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,12288,0.008968888885445064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,1024,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,6144,0.013071999781661563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,12288,0.01751377847459581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,16384,0.033835556772020124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,2560,0.006548444430033366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,1024,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,65536,0.05777155690722995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,64,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,6144,0.019681778219011094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,2560,0.049920890066358775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,2048,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,512,0.008272888759771982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,5120,0.017863111363516915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,16384,0.015815110670195687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,3584,0.021149333980348375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,12288,0.013421333498424955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,128,0.0028613333900769553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,64,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,4096,0.0502071115705702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,6144,0.022143110632896423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,6144,0.013829333086808523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,64,0.0028853333658642242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,8192,0.02148444453875224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,1024,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,12288,0.02439466615517934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,64,0.004584889031118817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,2048,0.0063484443558586975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,65536,0.12434755431281196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,128,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,8192,0.015636444091796875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,768,0.007260444263617198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,1024,0.005921777751710679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,2048,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,7168,0.012996444271670448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,1024,0.0070977773931291364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,768,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,256,0.008330666356616551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,1536,0.034726222356160484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,4096,0.00979288915793101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,6144,0.009869332942697737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,4096,0.01458666721979777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,2048,0.05071111188994514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,512,0.023746665981080797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,3072,0.01800355480776893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,65536,0.06514133347405328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,1024,0.006435555716355641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,3584,0.007345777418878343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,256,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,7168,0.014854222536087036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,2048,0.23316889339023164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,12288,0.3808346589406331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,5120,0.022589332527584497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,768,0.031126221021016438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,12288,0.01756622228357527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,2560,0.16646044784122044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,16384,0.0631200008922153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,3072,0.016526222229003906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,8192,0.015484443969196744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,128,0.0031991110493739447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,16384,0.02678577767478095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,32,0.0057546666098965546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,3584,0.05366488960054186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,10240,0.023615999354256526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,8192,0.020605333977275424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,2560,0.023394667439990576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,12288,0.006215999937719769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,65536,5.793504079182942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,32,0.005234666582610872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,8192,0.025098666548728943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,4096,0.017254221770498488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,2560,0.014857777290874057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,8192,0.024762666887707178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,3584,0.01110577748881446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,256,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,512,0.00489066665371259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,3584,0.010510221951537661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,64,0.002977777686383989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,6144,0.009071110851234859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,7168,0.010096888575288985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,10240,0.011864888999197217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,16384,0.026055110825432673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,128,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,2048,0.0070266665683852295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,5120,0.10108621915181477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,1536,0.011254222028785281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,2048,0.006624889042642381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,128,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,1536,0.02196888956758711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,10240,0.050252444214291044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,1024,0.23964089817470977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,128,0.006765333314736684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,128,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,2048,0.03700444433424208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,12288,0.009959110783206092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,7168,0.015793777174419828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,7168,0.017920888132519193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,768,0.016659556163681876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,512,0.024648000796635944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,128,0.029309332370758057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,16384,0.0481333335240682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,5120,0.008639111287064022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,2048,0.03832088907559713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,1536,0.023766222927305434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,1536,0.029698666599061754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,32,0.0028408887899584244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,7168,0.012445333103338877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,10240,0.03863999909824795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,3072,0.006792000184456508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,1536,0.009728888670603434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,5120,0.11694044537014431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,2048,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,512,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,2048,0.05260355605019463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,12288,0.013079110946920184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,7168,0.020076443751653034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,65536,0.0951111117998759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,256,0.016573333077960543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,2048,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,2560,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,3584,0.014255111416180929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,128,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,5120,0.018216000662909616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,4096,0.008571555217107138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,1024,0.013169777890046438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,128,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,1536,0.021659556362364028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,3584,0.03205422229237027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,3072,0.02010577751530541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,5120,0.1888959937625461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,16384,2.635786692301432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,3584,0.03138488862249587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,10240,0.01962577799956004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,7168,0.017216889394654166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,2048,0.01036888857682546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,65536,0.06483644247055054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,256,0.00833155545923445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,65536,0.13827733198801676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,32,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,1536,0.008407110969225565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,1024,0.004621333132187526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,1024,0.02401599950260586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,1024,0.025035555164019268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,1536,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,65536,0.2867208851708306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,5120,0.01851288808716668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,2560,0.00664444433318244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,7168,0.0670035547680325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,65536,0.12970044877794054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,32,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,6144,0.00944088896115621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,6144,0.03953866826163398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,4096,0.01789066692193349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,128,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,7168,0.05124533176422119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,5120,0.009024000002278222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,768,0.004231111043029361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,65536,0.16679555839962432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,4096,0.06839822398291694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,768,0.004239999999602635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,1536,0.03905955619282193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,256,0.005230222311284807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,32,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,128,0.0031226666437255014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,12288,0.01791466606987847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,12288,0.01720533271630605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,6144,0.024060444699393377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,6144,0.026901332868470088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,8192,0.028586665789286297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,1024,0.013753777576817406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,16384,0.06083822250366211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,8192,0.020921778347757127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,4096,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,2048,0.017836444907718234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,16384,0.030951112508773804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,128,0.015468445089128284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,768,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,768,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,8192,0.019513777560657926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,1536,0.027101332942644756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,12288,0.019549333386951022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,768,0.007251555720965068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,128,0.004852444347408083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,2048,0.06210844384299385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,8192,0.012738666600651212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,65536,0.044129778941472374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,7168,0.031799111101362444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,3072,0.025362667110231187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,6144,0.017510222064124215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,2048,0.011789333489206104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,3584,0.013239999612172445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,6144,0.017861333158281114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,32,0.009672889278994666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,128,0.003142222141226133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,2560,0.014169777433077494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,1024,0.007978666987684038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,64,0.0028284444577164124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,5120,0.008671999805503422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,3584,0.09195288684633042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,6144,0.06677244769202338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,65536,0.14447466532389322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,64,0.0028657778683635923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,1024,0.016195555528004963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,16384,0.01722311145729489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,2560,0.018200889229774475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,768,0.020536889632542927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,1536,0.004926222066084544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,65536,0.034104890293545194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,512,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,8192,0.013741333451535968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,256,0.014857777290874057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,7168,0.011758222348160215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,128,0.0028880000528362063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,8192,0.013106666505336761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,65536,0.13233333163791233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,256,0.003177777760558658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,6144,0.014492443866199918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,256,0.01926488843229082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,7168,0.007618666523032718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,128,0.0031644445326593188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,64,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,3072,0.009653333160612319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,5120,0.012521777715947894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,768,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,2560,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,32,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,10240,0.01720355616675483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,256,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,10240,0.029093331760830347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,32,0.0028568889117903183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,5120,0.00965244405799442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,16384,0.05611644188563029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,10240,0.015501333607567681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,128,0.014527999692493014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,6144,0.018352000249756705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,12288,0.23455378744337294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,3072,0.10393689076105754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,128,0.005236444373925527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,128,0.008969777988062965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,4096,0.013463999662134381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,2048,0.015127110812399121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,4096,0.0075866662793689305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,32,0.014196443888876172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,12288,0.2238453229268392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,4096,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,128,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,512,0.004602666530344221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,128,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,1024,0.034259554412629865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,1536,0.008257778154479133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,4096,0.013421333498424955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,10240,0.12461688783433701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,8192,0.12224088774787055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,8192,0.01772711177666982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,512,0.0041991112132867174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,3584,0.007664889097213745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,256,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,1024,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,7168,0.010787555740939246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,128,0.01239911135700014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,2560,0.013784888717863293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,12288,0.017245334055688646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,65536,0.07999289035797119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,1024,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,2560,0.14702310827043322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,8192,0.010597333312034607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,6144,0.025934222671720717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,12288,0.013456888496875763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,65536,0.051333334710862905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,4096,0.022962666220135156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,2560,0.02664888898531596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,4096,0.017364444004164804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,3072,0.006700444552633498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,7168,0.07215199867884318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,4096,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,2048,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,16384,0.046965334150526256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,65536,0.07446933454937406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,7168,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,512,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,3072,0.008543999658690559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,12288,0.12388088968065049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,1536,0.010404444403118556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,3584,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,10240,0.02701866626739502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,3072,0.012703110774358114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,3072,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,8192,0.10768800311618382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,1024,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,16384,0.03967555695109897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,32,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,4096,0.03549066517088149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,2560,0.006680888848172293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,3584,0.007358222372002072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,4096,0.010377777947319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,1536,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,8192,0.014515555567211576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,3072,0.028198222319285076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,65536,0.09676800171534221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,2560,0.01038933296998342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,512,0.0052355556852287715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,6144,0.01848088867134518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,6144,0.018274666534529794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,10240,0.20983555581834581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,128,0.020455999506844413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,10240,0.025144888295067683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,8192,0.020954666866196528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,2048,0.008101333346631791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,16384,0.03930044505331252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,128,0.007994666695594788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,768,0.004231111043029361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,64,0.0028195555011431375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,1536,0.10629244645436604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,2048,0.013554666605260638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,3584,0.030636443032158747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,128,0.01754133403301239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,65536,0.148181332482232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,65536,0.06934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,4096,0.11012355486551921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,65536,0.06416800287034777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,512,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,65536,0.09931644466188218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,6144,0.019871110717455547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,65536,0.08594222201241387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,65536,4.108562469482422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,7168,0.016206221448050607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,1536,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,3584,0.012507555385430654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,2048,0.0878408882353041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,512,0.005574222240183089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,16384,0.03307111064592997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,65536,0.08965777688556248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,5120,0.008982222113344405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,512,0.00886399961180157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,6144,0.014768888552983602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,16384,0.3086284531487359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,12288,0.03189155459403992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,3584,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,6144,0.02488800055450863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,5120,0.010017777482668558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,16384,0.03871999846564399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,1024,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,3072,0.014342221948835584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,768,0.021472889516088698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,65536,0.5190578036838108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,6144,0.2851626608106825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,12288,0.018219555417696636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,1536,0.016568000117937725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,16384,0.02607022225856781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,512,0.021993777818149988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,768,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,768,0.01684266659948561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,128,0.030692444907294378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,8192,0.010958222051461538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,65536,1.3062346776326497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,32,0.018177777528762817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,4096,0.02927644385231866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,12288,0.045447998576694064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,12288,0.00832622249921163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,3072,0.008761777646011775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,8192,0.013216888739003075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,5120,0.026492445005310908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,4096,0.012056888805495368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,2560,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,7168,0.027655111418830022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,2048,0.01586933268441094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,4096,0.014175999495718213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,2560,0.14606489075554743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,7168,0.03029600116941664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,128,0.010054222411579555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,3072,0.00721599989467197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,2560,0.007270221908887227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,16384,0.017151999804708693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,256,0.0031884445084465873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,3072,0.018894222047593858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,10240,0.026188444760110643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,1536,0.017438222964604694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,65536,0.05710577964782715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,7168,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,512,0.003206222214632564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,256,0.02270222206910451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,4096,0.031839112440745033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,6144,0.02249777813752492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,3072,0.0930817789501614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,128,0.0035377778112888336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,128,0.003064888839920362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,1536,0.012486221889654795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,10240,0.014148443937301636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,128,0.006763555523422029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,5120,0.009014222357008193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,2560,0.024996444582939148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,3072,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,512,0.0034728890491856467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,2560,0.16054844856262207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,16384,0.0244159996509552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,6144,0.015808888607554965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,3072,0.024111110303137038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,12288,0.0299893319606781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,2048,0.007345777418878343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,12288,0.02479733361138238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,2560,0.0209840006298489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,1024,0.005921777751710679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,128,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,6144,0.009355555805895064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,3584,0.08665155702167088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,32,0.005530666559934616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,8192,0.039488001002205744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,2048,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,7168,0.020396444532606337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,1536,0.02085688875781165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,1024,0.007738666401969061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,5120,0.05607910950978597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,256,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,256,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,7168,0.010050666828950247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,2048,0.011977777712874942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,2560,0.04342488778962029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,4096,0.015997333659066092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,32,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,5120,0.02162666618824005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,7168,0.018901333212852478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,12288,0.017791999710930716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,8192,0.03096710973315769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,4096,0.023212444451120164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,6144,0.01836266616980235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,768,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,128,0.01479288935661316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,1024,0.014458666245142618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,16384,0.1979831059773763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,7168,0.022982221510675218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,768,0.01702844434314304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,1024,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,1536,0.038097778956095375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,12288,0.029224889145957098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,256,0.020766221814685397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,256,0.0032097777972618737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,256,0.003244444520937072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,256,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,8192,0.08306044340133667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,4096,0.008043555749787224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,7168,0.018587556150224473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,512,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,7168,0.02320266597800785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,8192,0.13886133829752603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,5120,0.023546665906906128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,2048,0.016014221641752455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,12288,0.025412445267041523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,2048,0.025590222742822435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,8192,0.015558222929636637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,16384,0.06346044275495741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,64,0.0032382222513357797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,256,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,256,0.015069334043396844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,12288,0.06423999865849812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,7168,0.02812444501452976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,5120,0.24750399589538574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,12288,0.03457244568400913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,12288,0.014665777484575907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,16384,0.08074488904741076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,3072,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,32,0.0028195555011431375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,256,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,3584,0.007247111035717859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,6144,0.11680889129638672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,8192,0.010681777364677854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,3584,0.00794933322403166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,2560,0.023806222611003455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,3072,0.01255733354224099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,256,0.003826666623353958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,64,0.0032302221904198327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,256,0.005607111172543631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,128,0.0053102220926019884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,4096,0.009200889203283522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,16384,0.030903998348448012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,3584,0.017473777135213215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,4096,0.016884444488419425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,2048,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,256,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,12288,0.07647199763192071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,128,0.014492443866199918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,4096,0.010405333505736457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,2560,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,4096,0.014599111345079211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,65536,0.16870578130086264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,3584,0.25422665807935924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,10240,0.012080888781282636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,512,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,256,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,7168,0.03207644489076402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,512,0.0281333327293396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,3584,0.006225777582989798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,65536,0.11743822362687852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,3072,0.007339555356237624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,2048,0.010999110837777456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,16384,0.02035911050107744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,1536,0.007268444531493717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,2560,0.016156444946924847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,4096,0.047889779011408486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,65536,0.10008711285061306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,64,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,3072,0.19599910577138266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,1536,0.009015999734401703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,8192,0.06581955485873751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,768,0.011783999701340994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,65536,0.014081777797804939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,65536,0.31034665637546116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,512,0.013767999907334646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,2560,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,1024,0.011880888707107969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,8192,0.08159022198783027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,5120,0.02590222160021464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,32,0.0424560010433197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,8192,0.01590755581855774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,3072,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,4096,0.010679999987284342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,12288,0.04112889038191901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,7168,0.08781155612733628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,3072,0.10606488916609023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,10240,0.026816000541051228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,4096,0.03712088863054911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,7168,0.04012889001104567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,128,0.018076444665590923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,7168,0.02437866727511088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,3584,0.020090666082170274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,8192,0.01035911093155543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,65536,0.14684177769554987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,7168,0.023379556006855432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,7168,0.1152542167239719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,1536,0.01143200033240848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,3584,0.04725422130690681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,8192,0.046147555112838745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,6144,0.06735199689865112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,1536,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,7168,0.010439111126793755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,5120,0.03454666667514377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,12288,0.013694222602579327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,1536,0.017530666457282174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,4096,0.007592889169851939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,12288,0.04247022337383694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,512,0.0038844446341196695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,1024,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,6144,0.025643555654419795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,7168,0.029112001260121662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,16384,0.021716444028748408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,128,0.01516711049609714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,3072,0.015805333852767944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,65536,0.0636684430970086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,4096,0.13817689153883192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,256,0.0031235555393828284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,3584,0.0107351111041175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,512,0.021518222159809534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,512,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,6144,0.009364444348547194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,1536,0.03146222233772278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,16384,0.016505777835845947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,5120,0.019167110323905945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,7168,0.01759999990463257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,12288,0.009002666506502364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,8192,0.04359555575582716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,65536,0.1495057741800944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,65536,0.17973955472310385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,2560,0.0582631097899543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,256,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,7168,0.09197955661349827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,4096,0.06517422199249268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,5120,0.01514488955338796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,5120,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,2560,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,6144,0.032552000549104475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,16384,0.046624888976415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,3584,0.01074399964676963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,1024,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,512,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,256,0.015474667151769003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,3072,0.015807110402319167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,2560,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,256,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,2048,0.02844088938501146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,12288,0.01312711089849472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,8192,0.03442400031619602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,4096,0.008020444048775567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,3072,0.027115555273161993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,12288,0.033783111307356097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,5120,0.013571555415789286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,32,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,1024,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,6144,0.026698667142126296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,2048,0.008018666671382057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,512,0.0034799998005231223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,6144,0.15879555543263754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,3072,0.011055111057228513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,3072,0.026271111435360376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,2560,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,512,0.0033164444482988785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,16384,0.4122951030731201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,4096,0.017128000656763714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,7168,0.010207110808955299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,2048,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,8192,0.023585778143670823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,12288,0.1180915567609999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,1536,0.022257778379652236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,3584,0.010407110883129967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,7168,0.009671111073758867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,5120,0.012796444197495779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,65536,0.2527200116051568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,12288,0.26249244478013783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,3072,0.007994666695594788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,256,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,5120,0.27571378813849556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,1536,0.006608888920810487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,1536,0.008253333469231924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,6144,0.009648888475365108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,12288,0.02537155482504103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,1024,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,2048,0.015776889191733468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,3072,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,3072,0.007892444729804993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,256,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,256,0.0031679999083280563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,32,0.0025235555238193935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,256,0.004454222404294544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,256,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,2560,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,7168,0.013167111410035027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,512,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,12288,0.03489244315359328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,6144,0.010693333215183683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,5120,0.022388445006476507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,256,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,65536,0.06152977546056112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,7168,0.046168890264299184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,1024,0.01312888910373052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,3072,0.017234666479958426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,1536,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,3584,0.009391999906963771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,5120,0.09680977794859146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,10240,0.015492444237073263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,3072,0.014880888991885714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,7168,0.012157333393891653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,4096,0.03196355700492859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,6144,0.06738933589723375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,16384,0.03477777706252204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,65536,0.04896355668703715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,16384,0.02602311140961117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,3072,0.044440888696246676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,3584,0.007320889168315464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,10240,0.03442400031619602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,3584,3.959557427300347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,8192,0.04693688948949178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,2048,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,16384,0.08261599805619982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,3584,0.02604622145493825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,1024,0.06957599851820204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,256,0.02126666737927331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,10240,0.031055112679799397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,256,0.013663999736309052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,256,0.00793422261873881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,512,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,65536,0.16730754905276826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,1536,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,2048,0.008646222452322641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,1536,0.014835556348164877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,16384,0.07979733414120145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,3072,0.02699911097685496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,7168,0.010704889065689512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,2560,0.019242667489581637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,65536,0.08805244498782688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,64,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,16384,0.045363555351893105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,65536,0.12444533242119683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,1024,0.012031111452314587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,512,0.005127999931573868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,768,0.0038328886859946777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,1536,0.030496888690524634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,8192,0.010614222122563256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,3072,0.015809777710172866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,64,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,12288,0.023753777146339417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,4096,0.046690666013293795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,2048,0.040101332796944514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,10240,0.02197511163022783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,5120,0.016191111670600045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,3584,0.02794577678044637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,32,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,6144,0.019127999742825825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,5120,0.00851733320289188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,3584,0.01296622223324246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,512,0.008669333325492011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,512,0.007136888802051544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,2560,0.0196088883611891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,8192,0.011396444506115384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,128,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,1024,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,8192,0.01273155543539259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,768,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,3584,0.01350844403107961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,3584,0.015842666228612263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,32,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,64,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,12288,0.023017777336968318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,768,0.012356444365448423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,768,0.012545777691735161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,3584,0.012605333493815528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,1024,0.007671111159854465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,7168,0.057681779066721596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,8192,0.01124533348613315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,256,0.02568799919552273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,2560,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,1536,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,768,0.015465777781274585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,8192,0.022720000810093347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,2048,0.015519110692871941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,3072,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,128,0.005769777629110549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,256,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,3072,0.01051288925939136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,4096,0.10983288950390285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,16384,0.015775110986497667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,6144,0.013794666363133324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,8192,0.020643555455737643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,256,0.003216000066863166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,16384,0.04708177844683329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,5120,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,8192,0.13839111063215467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,256,0.023846222294701472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,7168,0.022343110707071092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,5120,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,256,0.045392000012927584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,512,0.18709066179063585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,7168,0.04158400164710151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,65536,0.03604799840185378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,64,0.005678222411208683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,2048,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,768,0.022395556171735127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,512,0.013419555293189155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,3584,0.009697777529557547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,12288,0.029333333174387615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,256,0.012828444441159567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,768,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,256,0.005558222118351195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,32,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,65536,0.06803733110427856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,16384,0.05800266398323906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,1024,0.009694221946928237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,12288,0.023318222827381555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,16384,0.03148533238304986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,4096,0.01976977785428365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,3072,0.016536000702116225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,10240,0.01626133256488376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,16384,0.02887111239963108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,512,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,8192,4.6560923258463545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,8192,0.03762311074468825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,1024,0.00777333312564426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,1536,0.017868444323539734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,2048,0.012153777811262341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,768,0.033928000264697604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,1024,0.03895111216439141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,5120,0.015480000111791821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,128,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,1024,0.006540444576077991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,2560,0.007190222541491191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,1024,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,1536,0.018215111560291715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,6144,0.0075795559419526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,4096,0.05594222413169014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,7168,0.010432889064153036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,768,0.010114666488435533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,3072,0.009209777745935652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,256,0.024364444944593642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,2048,0.006721777634488211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,16384,0.017198221551047433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,16384,0.12845422161950007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,16384,0.13759377267625597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,1024,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,10240,0.021617778473430212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,2048,0.007159111400445302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,256,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,3072,0.012353777885437012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,256,0.0035537779331207275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,1024,0.014734221829308404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,2560,0.007337777978844113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,12288,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,768,0.01720888912677765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,10240,0.018146667215559218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,10240,0.2563973267873128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,2560,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,1024,0.02383377816942003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,16384,0.023970666858885024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,128,0.0031431110368834603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,2560,0.021216000119845074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,1536,0.016522667474216886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,512,0.019254222512245178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,12288,0.01331911070479287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,6144,0.01976444489426083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,6144,0.230113771226671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,7168,0.013383111192120446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,3072,0.02296977738539378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,1024,0.004074666649103165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,16384,0.06034666962093777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,1024,0.0042453333735466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,512,0.0038542221817705366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,256,0.01275111072593265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,6144,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,1536,0.033618665403789945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,1024,0.03175910976197984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,16384,0.014440000057220459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,3072,0.008725333544943068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,1024,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,1536,0.004559999952713649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,3584,0.013768889009952545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,3584,0.0079297779334916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,768,0.020784000555674236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,2560,0.10243555572297837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,2048,0.006039111150635614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,16384,0.1106097764439053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,4096,0.04224444429079691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,768,0.027422222826215956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,64,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,2048,0.008000888758235509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,7168,0.03546133306291368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,10240,0.013332444760534497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,8192,0.010446222292052375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,10240,0.012553777959611682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,512,0.003519999898142285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,128,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,7168,0.07107644610934787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,1536,0.039953778187433876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,768,0.0035626664757728577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,256,0.010688888529936472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,2560,0.007261333366235097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,2560,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,512,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,128,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,8192,0.019144000278578866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,2560,0.015524443652894763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,16384,0.04198933309978909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,1024,0.031954665978749595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,10240,0.15425866179996067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,1024,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,1536,0.01015733347998725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,65536,0.15681866804758707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,7168,0.013646221823162503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,3072,0.010013333625263637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,4096,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,5120,0.007205333146784041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,12288,0.013076444466908773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,256,0.005558222118351195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,4096,0.02086755633354187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,768,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,256,0.004777777940034866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,12288,0.0413315561082628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,32,0.006978666616810693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,2560,0.006992888947327931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,12288,0.0336097776889801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,2560,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,1024,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,32,0.003108444520168834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,512,0.0038684445122877755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,1536,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,5120,0.010032888915803697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,7168,0.01038755559258991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,12288,0.10132710801230536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,32,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,4096,0.13180532720353869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,32,0.0032346666687064697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,1536,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,1536,0.00453599997692638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,12288,0.17954311105940077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,512,0.0033475555893447665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,2560,0.017265778448846605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,1024,0.003925333420435588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,10240,0.011781333221329583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,1536,0.00489777781897121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,1024,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,3584,0.00981866651111179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,1536,0.004875555634498596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,128,0.0031795555518733132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,5120,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,12288,0.014000889327791003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,4096,0.008000888758235509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,3072,0.019783111082182992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,2048,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,10240,0.03467822074890137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,12288,0.10352889034483169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,1536,0.004598222259018156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,5120,0.0547342234187656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,3584,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,3584,0.13546133041381836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,3584,0.007327111230956183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,2560,0.006913777854707506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,2560,0.011715555356608497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,2048,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,2560,0.010010666317409938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,128,0.0028524444335036804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,32,0.023165333602163527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,8192,0.023363555471102398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,512,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,3584,0.17336089081234404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,5120,0.05176000131501091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,3584,0.007473777565691207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,5120,0.00869955536392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,4096,0.012683555483818054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,1024,0.010677333507272931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,2560,0.016176889340082806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,2048,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,65536,0.10740533139970566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,7168,0.14656444390614828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,256,0.004243555582231945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,3584,0.007991111112965478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,1024,0.0038844446341196695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,2560,0.01477599971824222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,768,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,8192,0.05380355649524265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,2560,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,3584,0.01757600075668759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,1536,0.008903111020723978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,65536,0.08575555351045396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,2048,0.005535999933878581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,768,0.004558222161398994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,128,0.02030044462945726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,5120,0.012158222496509552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,16384,0.029077334536446467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,10240,0.03694933321740892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,768,0.011060444845093621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,1024,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,768,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,8192,0.023137778043746948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,1536,0.018947554959191214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,5120,0.018982221682866413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,5120,0.013419555293189155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,10240,0.011971555650234222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,2048,0.047037333250045776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,4096,0.11826577451494004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,512,0.015501333607567681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,2560,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,16384,0.047712001535627574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,16384,0.20131911171807182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,512,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,768,0.027849776877297297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,3072,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,32,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,5120,0.010603555374675326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,8192,0.011756444142924415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,768,0.016837333639462788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,8192,0.009336888790130615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,1024,0.00701155596309238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,2048,0.011315555208259158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,256,0.008336000144481659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,128,0.006624889042642381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,1024,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,12288,0.027150221996837195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,1024,0.024124445186720953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,4096,0.015499555402331881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,2048,0.014559111661381192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,32,0.003185777821474605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,128,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,128,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,32,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,8192,0.21355199813842773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,6144,0.025860443711280823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,3584,0.029024001624849107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,2048,0.00980444418059455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,16384,0.06023110946019491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,256,0.003896000070704354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,1024,0.02810133496920268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,7168,0.027553778555658128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,512,0.013435555828942193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,65536,0.08084977997673883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,10240,0.05065155691570706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,1536,0.012079999678664737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,7168,0.012830221818553077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,8192,0.010709332923094431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,8192,0.02927466564708286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,32,0.0028577778074476454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,12288,0.2846631209055583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,768,0.007070221834712558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,8192,0.014930667148696052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,3072,0.009776888622177972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,512,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,768,0.006568888823191325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,1024,0.016560888952679105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,1536,0.016891555653678048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,3072,0.021780444516075984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,768,0.02278933260175917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,768,0.00796088907453749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,3584,0.008628444539176093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,2560,0.008267555799749162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,12288,0.058584888776143394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,16384,0.06861688693364461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,65536,0.053447110785378345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,2048,0.008309333688682979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,6144,0.03376977642377218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,7168,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,1536,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,512,0.0031831111345026228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,12288,0.841682646009657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,2048,0.01127911110719045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,8192,0.05821422073576185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,2048,0.018244443668259513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,2048,0.048269334766599864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,3584,0.05671822362475925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,65536,0.5144737561543783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,32,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,5120,0.018968888454967074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,6144,0.023758222659428913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,4096,0.010444444086816577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,7168,0.024800888366169397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,65536,0.06408266888724433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,2560,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,10240,0.005222222043408289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,65536,0.14115644825829401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,5120,0.02513422237502204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,3072,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,65536,0.2645431094699436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,2048,0.13674222098456487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,12288,0.05457689033614265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,8192,0.14689332909054226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,256,0.003680888977315691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,12288,0.01975644462638431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,3072,0.021753778060277302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,512,0.004176889028814104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,65536,1.270776006910536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,7168,0.011443555355072021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,4096,0.016508445143699646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,6144,0.03635199864705404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,512,0.005662222289376789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,512,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,10240,0.02528711160024007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,512,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,5120,0.010131555298964182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,10240,0.034345779154035784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,6144,0.009073778159088558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,768,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,768,0.016229333149062265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,1536,0.004309333446953031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,128,0.005286222116814719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,128,0.002943111169669363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,12288,0.015483554866578845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,6144,0.3778968916998969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,2560,0.010100444157918295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,128,0.014831110835075378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,6144,0.008958222137557136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,3072,0.013668444421556262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,65536,0.1747413343853421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,12288,0.01774133410718706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,2048,0.03511910968356662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,64,0.003714666598372989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,7168,0.021261332763565913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,2560,0.016213332613309223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,65536,0.1404204501046075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,1024,0.016177778442700703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,16384,0.1572977834277683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,2048,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,16384,0.03105955653720432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,5120,0.25784889856974286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,768,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,5120,0.01145511120557785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,10240,0.01646222174167633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,8192,0.061944888697730176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,128,0.022876444790098403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,65536,0.06107288599014282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,2560,0.014977777997652689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,128,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,4096,0.00870666652917862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,2048,0.015522667103343539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,5120,0.014816888504558139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,16384,0.17005156146155465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,512,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,10240,0.018586667047606576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,256,0.014757333530320061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,2048,0.005197333378924264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,12288,0.025403555896547105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,512,0.02528000043498145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,16384,0.0317466656366984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,5120,0.008604444563388824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,3072,0.03068888849682278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,8192,0.030974222554100886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,256,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,512,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,512,0.006609777609507243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,16384,0.036179555786980525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,65536,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,1536,0.03605422377586365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,512,0.0031991110493739447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,4096,0.007616889145639207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,768,0.0041759999261962045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,16384,0.031126221021016438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,5120,0.008597333398130205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,768,0.0038577777643998465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,1024,0.020584000481499564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,512,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,65536,0.5792089038425022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,12288,0.2847653230031331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,3072,0.0169840008020401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,4096,0.017868444323539734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,3072,0.008361777497662438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,4096,0.025340444511837427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,65536,0.00999022192425198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,65536,0.014151111245155334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,2560,0.013408888545301227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,256,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,128,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,256,0.015459555718633862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,512,0.013464888764752282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,1536,0.326744900809394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,512,0.014503111441930136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,16384,0.015843555331230164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,7168,0.0123511114054256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,128,0.005552000055710475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,8192,0.02034133341577318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,3584,0.08020710945129395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,3584,0.01825066738658481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,768,0.009020444419648912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,65536,0.19852799839443633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,3584,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,4096,0.007662222617202335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,1024,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,32,0.0029235554652081597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,256,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,6144,0.03068444463941786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,256,0.00453244439429707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,768,0.016162667009565566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,8192,0.01814933286772834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,16384,0.015407111909654407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,6144,0.009316444396972656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,2560,0.0491555564933353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,32,0.0024862223201327855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,1536,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,3584,0.1595991055170695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,6144,0.010825778047243753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,5120,0.10050755739212036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,6144,0.18674133883582222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,8192,0.02055199940999349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,12288,0.030107554462220933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,12288,0.01757600075668759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,256,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,512,0.003847111016511917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,2560,0.007009777757856581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,3072,0.018552889426549275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,12288,0.016148444679048326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,256,0.01688799924320645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,1536,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,8192,0.021007999777793884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,5120,0.012759110993809171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,6144,0.04890399840142992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,128,0.014814222852389017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,12288,0.04993333419164022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,10240,0.041125333971447416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,32,0.011734222372372946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,7168,0.06722399923536512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,3072,0.007359111474619971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,65536,0.14000266128116184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,6144,0.15723289383782282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,64,0.006552888701359431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,768,0.00424177779091729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,1536,0.0063235556913746735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,5120,0.028970665401882593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,512,0.0035502223504914176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,3072,0.014185777968830533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,256,0.021966222259733412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,65536,0.6033964686923557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,3584,0.01033688916100396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,1536,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,2048,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,16384,0.022802667485343084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,3584,0.023689778314696416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,5120,0.019544000426928203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,6144,0.023296000228987798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,7168,0.009945777555306753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,3584,0.010325333310498131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,12288,0.4033679962158203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,3584,0.016224000189039443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,512,0.008280889027648503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,32,0.007244444555706448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,4096,0.08552889029184978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,1024,0.0404151115152571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,7168,0.12771733601888022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,16384,0.04011199871699015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,7168,0.022023111581802368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,7168,0.019131556153297424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,16384,0.2386488914489746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,3072,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,4096,0.011392888923486074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,4096,0.043990221288469106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,1024,0.0069102222720781965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,6144,0.10346488820181952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,2048,0.015222221612930298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,6144,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,4096,0.03629422187805176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,2560,0.015787555111779105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,256,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,8192,0.1054240001572503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,8192,0.01110400011142095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,3584,0.024666666984558105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,16384,0.015435554915004306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,3584,0.052858667241202466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,12288,0.05193066596984863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,1024,0.008399999803966945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,3072,0.009729777773221334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,2048,0.02828622195455763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,64,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,256,0.019298666053348117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,128,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,16384,0.032214221027162336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,3584,0.04228444563017952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,256,0.01516177753607432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,512,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,512,0.007657777931955125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,512,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,7168,0.011386666860845355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,512,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,1536,0.004603555632962121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,1536,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,256,0.018956444329685636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,6144,0.00963466697269016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,2560,0.016559110747443307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,16384,0.0703644421365526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,3072,0.017280000779363845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,512,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,4096,0.14989422427283391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,768,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,768,0.017286222842004564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,512,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,1024,0.007312888900438945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,1024,0.004234666625658671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,3584,0.010807111031479306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,256,0.017502221796247695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,2560,0.008295999632941352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,512,0.01555199921131134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,128,0.025389333566029865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,2560,0.0423351095782386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,4096,0.008007110820876228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,2048,0.07283288902706571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,2048,0.006592000110281839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,128,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,128,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,2560,0.009280000295903947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,5120,0.35047109921773273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,512,0.0035377778112888336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,1024,0.008316444853941599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,7168,0.010054222411579555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,10240,0.023063111636373732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,768,0.003657777690225177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,8192,0.023036445180575054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,768,0.01683022247420417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,3584,0.007640000018808577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,12288,0.18600355254279244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,256,0.01824977828396691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,64,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,1024,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,1024,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,12288,0.025037333369255066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,2048,0.02299022177855174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,8192,0.012991999586423239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,4096,0.010344888601038191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,3584,0.018933332628673978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,10240,0.017610667480362784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,5120,0.018945778409639995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,2048,0.00998488896422916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,12288,0.018400000201331247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,1536,0.005489777773618698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,8192,0.008027555214034187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,768,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,12288,0.027906666199366253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,10240,0.011909333368142446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,128,0.009815110928482479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,7168,0.019641778535313077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,8192,0.08132711384031507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,512,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,12288,0.18303199609120688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,4096,0.088755554623074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,512,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,3584,0.013707555830478668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,16384,0.13527111212412515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,16384,0.05454311105940077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,256,0.016944888565275405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,4096,0.008660444782839881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,65536,0.06407022476196289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,6144,0.029021332661310833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,7168,0.008341333104504479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,1024,0.03051466743151347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,1024,0.00369777778784434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,4096,0.020900444851981267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,128,0.0038808890514903595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,256,0.0147706667582194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,8192,0.028487109475665625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,768,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,256,0.00350222239891688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,12288,0.054261333412594266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,2560,0.009666666388511658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,32,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,7168,0.013369777964221107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,256,0.019155555301242404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,12288,0.04231111208597819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,12288,0.02050133380624983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,6144,0.01222400036123064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,7168,0.01203733351495531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,128,0.0035173334181308746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,3072,0.013833777772055732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,7168,0.03411644366052415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,128,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,1536,0.02334311107794444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,6144,0.01073066641887029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,7168,0.02235288918018341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,1536,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,32,0.012412444584899478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,10240,0.032617777585983276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,6144,0.009656888743241629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,10240,0.018035555879275005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,8192,0.17812355359395346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,2048,0.01623644431432088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,16384,0.02712000078625149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,6144,0.038698666625552706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,1536,0.0063324446479479475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,4096,2.339088863796658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,12288,0.07661955886416964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,16384,0.03921422362327576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,12288,0.018957333432303537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,32,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,1536,0.004502222355869082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,128,0.0028159999185138275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,5120,0.008966222405433655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,6144,0.022348443667093914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,1024,0.019001776973406475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,16384,0.0274311105410258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,128,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,4096,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,3584,0.022701332966486614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,64,0.02661866611904568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,5120,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,32,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,128,0.01500266624821557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,65536,0.10516622331407334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,6144,0.012396444049146442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,256,0.004234666625658671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,32,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,4096,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,12288,0.1087297797203064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,3584,0.04872533347871569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,64,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,5120,0.051184889343049794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,256,0.005234666582610872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,64,0.0038488888078265717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,4096,0.007696000238259633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,7168,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,1536,0.007979555262459649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,32,0.0031582222630580268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,256,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,5120,0.04340888725386726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,3072,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,16384,2.011167950100369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,4096,0.016858667135238647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,5120,0.008962666822804345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,128,0.004171555654870139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,10240,0.030234667989942763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,3072,0.03089244498146905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,7168,0.010030222435792288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,256,0.0069520001610120135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,8192,0.0248346659872267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,256,0.005695111221737332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,1536,0.02600266701645321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,16384,0.02239466706911723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,16384,0.01570133368174235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,2048,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,64,0.018720888429217868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,128,0.0030586665703190696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,256,0.017145777742067974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,6144,0.016303999556435477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,10240,0.015575110912322998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,768,0.0041786668201287585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,1024,0.04370222157902188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,6144,0.12018133534325494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,2048,0.0052737775776121355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,1024,0.02312622302108341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,768,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,64,0.0029084444459941653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,128,0.020968889196713764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,768,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,768,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,10240,0.025921778546439275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,16384,0.08416799704233806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,65536,0.012726222475369772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,2048,0.02093066606256697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,8192,0.01372977760103014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,1536,0.009308444129096137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,768,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,1024,0.01314933349688848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,2048,0.008985777695973715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,65536,0.14923732810550266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,4096,0.7094782193501791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,3072,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,256,0.012411555482281579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,10240,0.01445777714252472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,12288,0.03050400151146783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,8192,0.03264800045225356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,16384,0.015407111909654407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,7168,0.014462222655614218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,4096,0.02164977788925171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,2048,0.02336799932850732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,12288,0.01936533384852939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,12288,0.08247111241022746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,128,0.01372977760103014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,8192,0.05693778064515856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,5120,0.12794933054182264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,6144,0.019387554791238572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,512,0.0069120000633928515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,2560,0.008303110798199972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,3584,0.018228444788191054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,12288,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,3584,0.038712888956069946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,128,0.0039013334446483185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,6144,0.01715377800994449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,1024,0.02718577782313029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,12288,0.05338933401637607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,2048,0.030424889591005113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,2560,0.014477334088749357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,65536,0.06807555754979451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,16384,0.06824888785680135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,3072,0.011720000041855706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,65536,0.19053244590759277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,256,0.0031644445326593188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,768,0.0038968887594011095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,64,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,7168,0.013965333501497904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,12288,0.014448000325096978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,16384,0.01642311116059621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,32,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,64,0.0028106667515304354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,3584,0.010299555957317352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,256,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,12288,0.028755555550257366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,8192,0.02441155579355028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,6144,0.00963644435008367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,5120,0.11099200116263495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,512,0.01148799972401725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,3072,0.007303111255168915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,3584,0.010069333016872406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,512,0.005320888840489917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,3072,0.01586844523747762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,1024,0.02102488941616482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,256,0.016566221912701923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,3584,0.06917422347598605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,2560,0.006955555743641323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,3072,0.0075902218619982404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,8192,0.012157333393891653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,4096,0.04214310977194044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,1024,0.014464888307783337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,2560,0.06274666388829549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,65536,0.14657599396175808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,32,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,3584,0.007991111112965478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,10240,0.017180444465743173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,2560,0.01685688893000285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,1024,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,32,0.00315022220214208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,2048,0.006573333508438534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,2048,0.015823110938072205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,512,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,32,0.005320000151793162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,8192,0.01512977812025282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,6144,0.04819555415047539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,1024,0.004262222184075249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,1024,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,1536,0.004557333472702238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,7168,0.030221333106358845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,2560,0.010713777608341642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,2048,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,6144,0.02342044479317135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,768,0.009190222455395592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,6144,0.011620444556077322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,4096,0.012910222013791403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,7168,0.023176888624827068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,1024,0.016075554821226332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,4096,0.013175999952687157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,4096,0.009992000129487779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,1024,0.06191200017929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,512,0.007245333658324347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,3072,0.012149333126015134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,7168,0.1377484401067098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,6144,0.007244444555706448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,16384,0.016198222835858662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,5120,0.01884622209601932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,8192,0.052665776676601835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,8192,0.011457777685589261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,3072,0.018763555420769584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,128,0.009687110781669617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,16384,0.05793688694636027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,3584,0.010749333434634738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,7168,0.018254222141371835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,256,0.03625955515437656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,2048,0.008634666601816813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,2560,0.015876443849669564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,64,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,32,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,3072,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,128,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,6144,0.013981334037250943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,12288,0.09753244453006321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,256,0.00554577757914861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,16384,0.02490577763981289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,2560,0.007556444240940942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,7168,0.013386666774749756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,256,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,10240,0.017119111286269296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,8192,0.01072533345884747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,1024,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,8192,5.263167063395183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,7168,0.07837955819235908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,16384,0.03496799866358439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,64,0.004586666822433472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,256,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,65536,0.10415911012225681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,2560,0.24574844042460123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,7168,0.04629866613282097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,8192,0.02693066663212246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,16384,0.016173332929611206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,10240,0.021373333202468023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,12288,0.02570844524436527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,64,0.009471110999584198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,10240,0.022708444131745234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,3584,0.01647644407219357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,128,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,10240,0.011310222248236338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,65536,0.14079911179012722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,768,0.006933333145247565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,2560,0.011237333218256632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,5120,0.011015111373530494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,8192,0.12656799952189127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,2560,0.00795288880666097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,6144,0.018553778529167175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,16384,0.022307554880777996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,4096,0.008284444610277811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,7168,0.05314755439758301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,3072,0.010328888893127441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,768,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,32,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,3584,0.01919644408755832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,5120,0.1342888938056098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,1536,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,4096,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,64,0.003167111012670729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,12288,0.03341955608791775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,64,0.0028737777223189673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,8192,0.016880000631014507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,3584,0.012563555604881711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,4096,0.010570666856235929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,4096,0.00794666674402025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,3072,0.011043555206722684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,1024,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,2560,0.012320888539155325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,32,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,768,0.003955555458863576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,6144,0.017264000243610807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,3584,0.0350195566813151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,64,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,3584,0.012803555362754397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,3584,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,3584,0.01257599973016315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,1536,0.032503998941845365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,768,0.008847999903890822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,32,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,16384,0.09310133589638604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,16384,0.04378133349948459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,1024,0.0070062221752272705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,768,0.015505777464972602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,2560,0.011896889242861005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,7168,0.129622220993042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,2048,0.006677333265542984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,2048,0.018586667047606576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,12288,0.3984231154123942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,3584,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,256,0.012821333275900947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,65536,0.09081688854429458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,2048,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,128,0.00665511108107037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,65536,0.05920266442828708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,7168,0.06515822145673963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,65536,0.15009600586361355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,4096,0.007633777956167857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,12288,0.07649244202507867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,65536,0.14865510993533665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,3072,0.007749333149856991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,7168,0.011439111497667102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,256,0.009006222089131674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,16384,0.020426667398876615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,12288,0.02772177755832672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,3072,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,768,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,7168,0.057393776045905225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,4096,0.04993599984380934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,8192,0.02884888980123732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,7168,0.011406222151385413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,16384,0.10062844223446316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,32,0.004912888838185204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,128,0.015479111009173922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,1024,0.0052177777720822235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,3072,0.11302399635314941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,8192,0.0446284446451399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,1024,0.011536888778209686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,512,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,3584,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,4096,0.02310488952530755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,4096,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,6144,0.00963644435008367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,7168,0.10419733656777276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,64,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,10240,0.13098577658335367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,10240,0.03219288918707106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,7168,4.181972503662109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,5120,0.019321777754359774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,768,0.011752888560295105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,16384,0.045893334680133395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,6144,0.03912711143493652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,512,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,2560,0.006945778098371293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,8192,0.03197066651450263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,1536,0.01480977733929952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,6144,0.013749333719412485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,8192,0.6614693535698785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,3072,0.009337777892748514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,3072,0.010332444475756751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,10240,0.16374844974941677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,4096,0.25331200493706596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,65536,0.20210133658515084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,2048,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,65536,0.1863457759221395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,12288,0.011672000090281168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,1536,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,256,0.006903111106819577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,32,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,8192,0.07578844494289823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,3072,0.050026665131251015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,64,0.015903110305468243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,10240,0.023205333285861548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,10240,0.26003289222717285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,1536,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,6144,0.030594666798909504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,5120,0.018522666560279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,6144,0.05288711190223694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,65536,0.22233332528008354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,128,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,3584,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,128,0.003510222252872255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,32,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,2560,0.28008000055948895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,1024,0.006219555520349079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,65536,0.17034667068057588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,768,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,4096,0.007662222617202335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,7168,0.00904266701804267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,2560,0.09433599975374009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,10240,0.040272000763151385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,10240,0.018363555272420246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,3072,0.009735110733244155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,3072,0.007634667058785756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,32,0.0041742221348815495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,2560,0.012746666868527731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,32,0.003168888803985384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,512,0.029325332906511094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,128,0.013821333646774292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,10240,0.027112889620992873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,7168,0.03128355410363939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,16384,0.02305511136849721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,16384,0.03734044565094842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,12288,0.04514133267932468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,1024,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,2560,0.009349333743254343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,768,0.008216000265545314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,512,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,4096,0.017498667041460674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,8192,0.09162044525146484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,2048,0.007124444676770105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,64,0.006882666713661618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,3584,0.00960533320903778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,1024,0.021262221866183813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,4096,0.01112711098459032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,128,0.015124445160230001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,5120,0.08608000146018134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,128,0.014112000664075216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,2048,0.0967786643240187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,65536,0.14540533224741617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,4096,0.007974222302436829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,2560,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,2048,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,512,0.011757333245542316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,8192,0.025238222546047632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,2048,0.009700444837411245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,512,0.00871288859181934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,8192,0.021370665894614324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,8192,0.03949688871701559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,16384,0.023415999280081853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,2560,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,5120,0.22195111380683052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,5120,0.007531555162535773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,2048,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,768,0.015476443701320224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,1536,0.06711822085910372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,16384,0.02849688794877794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,12288,0.0492142207092709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,4096,0.02647200061215295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,4096,0.00903733323017756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,3072,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,1024,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,1024,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,4096,0.0127688886390792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,12288,0.12989955478244358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,3584,0.014861333701345654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,2048,0.007247111035717859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,2560,0.10134666495853001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,5120,0.00830577810605367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,768,0.004182222402758068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,2560,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,1536,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,16384,0.03413688805368211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,1536,0.006304889089531369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,3072,0.00795733349190818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,128,0.0031662223239739737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,16384,0.06390488810009427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,3584,0.01888088881969452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,2048,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,7168,0.16448977258470324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,16384,0.015446222490734525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,8192,0.03583555420239767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,7168,0.02812622321976556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,7168,0.010234666367371878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,5120,0.012441777520709567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,3072,0.0075937774446275495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,10240,0.04355466696951124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,3584,0.014926221635606555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,16384,0.02055911057525211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,2560,0.043931557072533496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,256,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,128,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,1536,0.013814222481515674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,5120,0.02715911136733161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,3584,0.02180266711446974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,768,0.03567911187807719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,1536,0.01274400038851632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,2048,0.1118239959081014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,6144,0.012104888757069906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,12288,0.02272533377011617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,768,0.0551457769340939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,2048,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,2560,1.404520034790039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,4096,0.057778669728173145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,65536,0.08258666594823201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,2560,0.014140443669425117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,65536,0.1974568896823459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,3072,0.007680888805124495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,4096,0.019480889042218525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,32,0.0031386667655573953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,2560,0.014845333165592618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,1536,0.01015377789735794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,4096,0.012537777423858643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,3584,0.009738666315873465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,2048,0.012771555946932899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,7168,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,1536,0.01549600064754486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,10240,0.025451555848121643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,768,0.003499555504984326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,3072,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,256,0.007633777956167857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,16384,0.03696889016363356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,1024,0.0155404441886478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,768,0.017272889614105225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,256,0.005883555445406172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,512,0.00612088872326745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,128,0.01037688884470198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,4096,0.007826666865083907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,64,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,768,0.00462311092350218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,64,0.005973333285914526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,64,0.005088888936572605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,512,0.020626667473051283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,65536,0.0843413339720832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,2048,0.015855999456511605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,5120,0.025756445195939805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,2560,0.007542221910423703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,2048,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,8192,0.012183999849690331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,256,0.016512889001104567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,768,0.007595555649863348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,4096,0.05125511023733351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,256,0.014472888575659858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,1536,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,5120,0.00794400026400884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,768,0.021128889587190416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,8192,0.048379557000266186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,2048,0.006172444257471297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,8192,0.03574666712019179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,5120,0.009031111167536842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,2048,0.03278399838341607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,8192,0.2040799988640679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,1536,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,128,0.00453955555955569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,12288,0.17446045080820718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,5120,0.02269777821169959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,3584,0.007330666813585494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,12288,0.027610666222042505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,12288,0.01605422298113505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,6144,0.03237511052025689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,7168,0.00999466660949919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,5120,0.023119111855824787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,128,0.006002666635645761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,65536,0.4848053190443251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,4096,0.008302222523424361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,2560,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,4096,0.005210666606823604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,6144,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,2048,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,1024,0.013416888813177744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,64,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,64,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,5120,0.018613333503405254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,8192,0.03108711043993632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,768,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,65536,0.23672355545891655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,5120,0.1292862229877048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,64,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,4096,0.007920888562997183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,6144,0.012730666332774691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,64,0.0028648889727062653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,2560,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,2048,0.006924444602595435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,512,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,4096,0.012800888882742988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,1024,0.011035555766688453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,512,0.03384711013899909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,7168,0.024870221813519795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,10240,0.03645599881807963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,4096,0.016149333781666227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,128,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,4096,0.017316444052590262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,256,0.014853333433469137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,768,0.028359999259312946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,7168,0.01646933290693495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,1024,0.015808888607554965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,7168,0.010042666561073726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,10240,0.02276977731121911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,7168,0.021425777011447485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,1024,0.007231111327807109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,512,0.013436444103717804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,10240,0.021976888179779053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,3072,0.28428623411390513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,768,0.012412444584899478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,3584,0.00902844468752543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,32,0.0041271112859249115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,10240,0.02919911013709174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,5120,0.008661333057615492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,4096,0.021689777572949726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,16384,0.024332443873087566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,65536,0.08236977789137098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,12288,0.23691466119554308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,12288,0.020399999287393358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,1024,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,256,0.0031519999934567344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,3584,0.027855998939938013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,768,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,1536,0.006217777729034424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,6144,0.009305777649084726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,7168,0.026143999563323125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,3072,0.007281777759393056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,5120,0.008007110820876228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,12288,0.013108444710572561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,512,0.005692444327804778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,16384,0.06208178069856432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,3584,0.05029955506324768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,768,0.0189057770702574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,4096,0.021329777108298406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,10240,0.057497779528299965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,10240,0.040320889817343816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,12288,0.024803555674023096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,4096,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,256,0.010746666954623329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,64,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,6144,0.014797333214018079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,64,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,65536,0.062246223290761314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,4096,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,2560,0.022287999590237934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,4096,0.00998488896422916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,6144,0.4358551237318251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,256,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,12288,0.09226133426030476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,12288,0.2958035469055176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,10240,0.01070666644308302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,10240,0.01609422266483307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,32,0.00723555518521203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,128,0.02197244432237413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,2560,0.26483021842108834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,768,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,16384,0.03515733281771342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,256,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,1024,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,1536,0.014478221535682678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,6144,0.0088746663596895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,512,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,32,0.0032088889016045462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,8192,0.03525244527392917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,6144,0.27541955312093097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,256,0.020635555187861126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,2048,0.05294933252864414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,768,0.006350222147173352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,6144,0.023989333046807185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,2560,0.025461332665549383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,3584,0.008366222182909647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,12288,0.03272799981964959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,2048,0.028365333875020344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,2560,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,3584,0.007670222057236566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,768,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,1024,0.012402666939629449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,128,0.026224000586403742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,12288,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,7168,0.0116462219092581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,768,0.037054220835367836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,1024,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,4096,0.007930667036109501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,4096,0.008243555823961893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,1536,0.007810666329330868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,256,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,128,0.05385422375467089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,1536,0.024132443798912898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,768,0.0036142224238978494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,4096,0.020436444216304354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,1024,0.024314666787783306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,2048,0.01443733274936676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,12288,0.3205671045515272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,10240,0.03557066784964667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,1024,0.006676444576846228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,256,0.023036445180575054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,512,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,6144,0.018806222412321303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,16384,0.038482666015625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,7168,0.021918222308158875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,256,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,4096,0.01718844473361969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,512,0.012144000165992312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,7168,0.01924000018172794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,2048,0.01648888819747501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,65536,0.07525777816772461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,2048,0.023472888602150813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,768,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,6144,0.035192890299691096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,256,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,768,0.005554666535721884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,3584,0.09276622533798218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,3584,0.008018666671382057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,12288,0.052384889788097806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,5120,0.01719911065366533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,7168,0.031050665511025324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,32,0.17220444149441191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,256,0.0028328889360030494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,10240,0.28719109959072536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,10240,0.04564177658822802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,128,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,512,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,3584,0.014863111906581454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,256,0.004550222307443619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,128,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,5120,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,3584,0.05396355523003472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,7168,0.019489778412712943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,1536,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,3584,0.023784889115227595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,3072,0.013547555440002017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,5120,0.07153155406316121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,64,0.002903999967707528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,512,0.006025777922736273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,12288,0.014555555250909595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,2560,0.3605964448716905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,128,0.007371555599901412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,768,0.01795111099878947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,7168,0.04197866717974345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,10240,0.10753333568572998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,5120,0.052575111389160156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,3584,0.015519110692871941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,32,0.02583644456333584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,128,0.0031235555393828284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,3072,0.06780622402826945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,768,0.003522666792074839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,7168,0.038980444272359215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,768,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,3584,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,1536,0.004507555729813046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,3072,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,8192,0.013134222063753339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,10240,0.016394666499561734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,10240,0.04055466585689121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,3072,0.015806222955385845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,12288,0.03598399957021078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,5120,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,12288,0.018007111218240526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,6144,0.03003022074699402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,256,0.005828444328573015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,2560,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,12288,0.024728889266649883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,1024,0.0058471109304163195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,65536,0.3443137804667155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,512,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,1536,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,3584,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,3072,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,768,0.037032001548343234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,1024,0.05512711074617174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,1536,0.009955555200576782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,16384,0.06343555450439453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,64,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,10240,0.15411467022365993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,1024,0.005940444353553984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,5120,0.03931644558906555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,512,0.003571555432346132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,12288,0.06007822354634603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,4096,0.005594666633341048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,16384,0.016131555040677387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,3072,0.02141422198878394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,16384,0.04425244530042013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,1536,0.017539555827776592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,4096,0.07917333311504789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,65536,1.2351581785413954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,6144,0.021930666433440313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,7168,0.019739554988013375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,1024,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,256,0.005207111024194294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,512,0.06472533278995089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,64,0.006593777901596493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,12288,0.030252443419562444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,2560,0.015860444969601102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,2560,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,12288,0.061556445227728955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,12288,0.01311644415060679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,6144,0.015678221980730694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,2560,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,3584,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,3584,0.015795555379655626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,32,0.007338666253619724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,512,0.005980444451173146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,1536,0.03803733322355483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,5120,0.013746666411558786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,5120,0.016539555456903245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,1024,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,3072,0.015535111228624979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,3072,0.016879111528396606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,512,0.01904266741540697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,2048,0.011703111231327057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,32,0.003536888708670934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,64,0.0031759999692440033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,64,0.002846222163902389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,4096,0.016516443755891588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,1024,0.01754133403301239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,128,0.002830222249031067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,12288,0.3891039954291449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,7168,0.01519466605451372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,1536,0.016864889197879367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,128,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,3072,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,16384,0.06836622291141085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,10240,0.01145066652033064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,128,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,5120,0.02384177843729655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,6144,0.018242667118708294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,1024,0.02662044432428148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,3584,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,128,0.01609244445959727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,7168,0.03924622138341268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,2560,0.008358221915033128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,65536,0.17008088694678414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,512,0.006072000082996156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,4096,0.01771022213829888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,6144,0.017229333519935608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,65536,0.07251644134521484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,1024,0.007324444750944774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,4096,0.01977244516213735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,6144,0.0229413327243593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,10240,0.9259075588650174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,1024,0.02027466727627648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,512,0.01963555481698778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,1024,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,768,0.041484445333480835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,3072,0.012830221818553077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,8192,0.02970400121476915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,5120,0.013844444519943662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,6144,0.01809244520134396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,7168,0.007965332931942409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,1536,0.008988444175985124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,10240,0.09229244126213922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,8192,0.09383022122912937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,6144,0.027875555886162653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,4096,0.007989333735571967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,1024,0.05577511257595486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,6144,0.009473777479595607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,1536,0.008665777742862701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,65536,0.06135466363694933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,12288,0.09687822394900852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,1024,0.008216888540320927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,768,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,3072,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,4096,0.017321777012613084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,2048,0.014515555567211576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,10240,0.020117332537968952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,3584,0.02255288925435808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,16384,0.6825902197096082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,4096,0.016143111719025504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,3584,0.018181333939234417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,256,0.01886577738655938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,128,0.003463111196955045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,256,0.013646221823162503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,3072,0.007256000406212277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,65536,0.10909244749281143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,4096,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,8192,0.027408000495698717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,6144,0.00755555596616533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,5120,0.01754133403301239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,256,0.015806222955385845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,8192,0.011214222345087262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,6144,0.02994311187002394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,2048,0.008623999853928884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,12288,0.018534221582942538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,256,0.004494222087992562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,4096,0.012063999970753988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,8192,0.02830044428507487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,2560,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,2560,0.014031999640994601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,8192,0.014307555225160388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,128,0.014067555467287699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,512,0.010029333333174387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,768,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,6144,0.009184000392754873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,2560,0.03473866648144192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,128,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,128,0.015180443723996481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,768,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,12288,0.02442399991883172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,2560,0.022109333011839125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,768,0.004815111143721474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,16384,0.035411556561787925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,7168,0.017525333497259352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,65536,0.590213351779514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,2048,0.016192888220151264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,256,0.005535111245181825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,32,0.004544888933499654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,128,0.01621955633163452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,2048,0.011749332977665795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,3584,0.041907555527157254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,7168,0.12807377179463705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,2560,0.00654666663871871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,8192,0.028576887316173975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,5120,0.05895644426345825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,64,0.0027884443600972495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,65536,0.13711378309461805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,512,0.007258666886223688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,12288,0.0365573333369361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,512,0.02574399941497379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,10240,0.14587822225358751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,3584,0.008360889222886827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,16384,0.017880888448821176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,768,0.00423911131090588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,256,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,768,0.015429332852363586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,65536,0.259152889251709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,1536,0.008240888516108194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,512,0.0058755555914507965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,8192,0.9025564193725586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,12288,0.4181413385603163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,8192,0.024710221423043147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,65536,0.0627982219060262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,12288,0.013980444934633044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,3072,0.01068800025516086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,4096,0.010412444670995077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,16384,0.017257778180970084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,10240,0.012533333566453723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,3584,0.008999110923873054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,16384,0.06931377781762017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,4096,0.022686221533351477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,2560,0.01479644411140018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,4096,0.008064000142945183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,64,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,7168,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,16384,0.04896622233920627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,6144,0.012878222597969903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,1536,0.006969778074158563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,7168,0.012179555164443122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,1024,0.014879999889267815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,1536,0.05083466569582621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,256,0.015246222416559855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,16384,0.23262755076090494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,64,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,65536,0.07347555292977227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,16384,0.12045866913265652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,65536,0.4220248858133952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,512,0.008976889153321585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,12288,0.5488053427802192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,7168,0.010690666735172272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,7168,0.0147706667582194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,1536,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,256,0.006559999866618051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,768,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,6144,0.035070220629374184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,2560,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,1536,0.010500444306267632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,4096,0.0380639996793535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,6144,0.02235377828280131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,256,0.027671999401516382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,3072,0.008615111311276754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,1024,0.01724177764521705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,8192,0.014131555954615274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,2560,0.008060444560315874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,3072,0.009657777845859528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,4096,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,1024,0.06128622425927056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,128,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,512,0.004107555581463708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,7168,0.014167999227841696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,16384,0.016978666186332703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,128,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,1536,0.016167110866970487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,5120,0.006232888748248418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,8192,0.011391110718250275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,8192,0.5214382277594672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,2560,0.012671111358536614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,1024,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,512,0.004601777841647466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,5120,0.018457776970333524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,12288,0.04394933250215319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,2560,0.00867377801073922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,256,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,1536,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,65536,0.1375982231563992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,64,0.0031306667046414483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,4096,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,4096,0.008664888640244802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,5120,0.036609778801600136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,6144,0.012200888660218982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,128,0.006835555450783835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,512,0.006913777854707506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,4096,0.014067555467287699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,7168,0.02534666657447815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,256,0.007648000286685095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,1536,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,1024,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,1536,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,8192,0.02076088885466258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,4096,0.02204533252451155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,256,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,6144,0.030045333835813735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,4096,0.007978666987684038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,32,0.006189333481921091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,5120,0.006416000011894438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,65536,0.06932977835337321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,10240,0.03751466671625773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,2560,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,1024,0.00526311124364535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,256,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,2560,0.013353777428468069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,1536,0.014107555150985718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,65536,0.072580443488227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,12288,0.02881422307756212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,3072,0.06109155548943413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,65536,0.07821777794096205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,512,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,4096,0.09771288765801324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,6144,0.08444711234834458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,768,0.0038364442686239877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,65536,0.07595911290910509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,7168,0.01000533335738712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,2048,0.007993777592976889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,512,0.0031946665710873077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,16384,0.015481778317027621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,256,0.003162666741344664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,3584,0.008436444732877944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,2048,0.020766221814685397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,12288,0.019715555840068393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,128,0.0064071110553211635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,6144,0.02162666618824005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,65536,5.528502146402995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,256,0.009988444546858469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,12288,0.044700443744659424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,768,0.03286666671435038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,256,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,7168,0.01663377715481652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,16384,0.04636444316969978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,4096,0.008283555507659912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,12288,0.01371911085314221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,3584,0.01482222146458096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,2048,0.008289777570300633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,768,0.016178665889634024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,128,0.005196444276306364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,10240,0.02789688772625393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,128,0.019111111760139465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,10240,0.019356444478034973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,512,0.011568889021873474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,6144,0.14793422487046984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,512,0.0052800000541739995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,2560,0.03201333350605435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,10240,0.016146666473812528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,12288,0.036908444431093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,128,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,10240,0.025736000802781846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,512,0.007994666695594788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,1024,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,256,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,65536,0.207750227716234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,10240,0.017621333400408428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,3072,0.01867999964290195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,3584,0.013108444710572561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,1536,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,12288,0.04627644353442722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,64,0.0028977776981062363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,5120,0.018840889135996502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,768,0.005691555639108022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,16384,0.02056711084312863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,6144,0.01239911135700014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,512,0.026502221822738647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,128,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,1024,0.007304888632562425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,6144,0.012109333442317115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,10240,0.010018666585286459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,512,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,12288,0.04327999883227878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,65536,0.010069333016872406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,768,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,12288,0.02238044473859999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,16384,0.0465191106001536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,128,0.008247999681366814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,3584,0.00797333319981893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,3584,0.010703110860453712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,4096,0.01942133406798045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,3584,0.007589333587222629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,32,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,5120,0.02409511142306858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,256,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,768,0.007897777689827813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,64,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,16384,0.05761155817243788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,768,0.007624000310897827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,128,0.0031848889258172777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,256,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,3584,0.03647733396954007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,6144,0.009769777456919352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,128,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,4096,0.02030755579471588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,65536,0.191631105211046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,1536,0.006874666859706243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,16384,0.02358755634890662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,12288,0.03107822272512648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,2560,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,7168,0.020416888925764296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,4096,0.023187556200557288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,3072,0.05964533487955729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,1536,0.1507937775717841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,12288,0.013434666726324292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,10240,0.016280000408490498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,12288,0.4998764461941189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,4096,0.013145777914259167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,1024,0.010858666565683154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,64,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,2048,0.019905777441130746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,5120,0.012103111379676394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,10240,0.15892444716559515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,16384,0.023780445257822674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,6144,0.053120000494851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,7168,0.03019022279315525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,6144,0.012794666820102267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,12288,0.042103999190860324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,8192,0.02242044442229801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,768,0.0058755555914507965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,3584,0.016191111670600045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,1024,0.03403110967742072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,12288,0.029804444975323145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,768,0.015206222732861837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,1024,0.018256889449225534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,7168,0.010416888528399997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,2560,0.11853599548339844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,64,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,65536,0.08182311058044434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,10240,0.02784444557295905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,7168,0.009658666948477427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,64,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,4096,0.017872888180944655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,1536,0.32802311579386395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,8192,0.027472888429959614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,1024,0.004182222402758068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,5120,0.01039644413524204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,5120,0.018225777480337355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,768,0.016700444949997794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,7168,0.016306666864289176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,2048,0.013804444836245643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,256,0.005396444350481033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,12288,0.013459555804729462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,512,0.013305777476893531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,10240,0.015470221638679504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,512,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,6144,0.02163999941613939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,1024,0.007631111476156447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,1024,0.003879111260175705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,12288,0.05102666550212436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,5120,0.0735608869128757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,6144,0.012513777448071374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,12288,0.013391999734772576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,256,0.02608888844648997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,768,0.02575911084810893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,2560,0.017238222890430026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,8192,0.011030221978823343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,5120,0.013080889152155982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,1536,0.006909333169460297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,7168,0.010030222435792288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,3584,0.03354310989379883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,16384,0.10216444730758667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,7168,0.013150222599506378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,2560,0.9143377939860026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,8192,0.01886666648917728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,12288,0.11846310562557644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,3072,0.035840001371171736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,4096,0.014860444598727755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,64,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,5120,0.016209777858522203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,4096,0.016568889220555622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,64,0.006606222026877933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,768,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,1536,0.006552888701359431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,1536,0.006148444281684027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,8192,0.029439111550649006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,3584,0.014149333039919535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,6144,0.020644444558355544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,3584,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,7168,0.00960000024901496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,256,0.02126844392882453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,3584,0.011054221954610614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,4096,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,4096,0.007711999946170383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,2560,0.010007999837398529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,1024,0.014834667245546976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,128,0.016228444046444364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,5120,0.03767111235194736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,3072,0.007048889166778988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,6144,0.023987554841571387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,3072,0.011368888947698804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,6144,0.03251911203066508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,16384,0.06927377647823758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,2048,0.026016000244352553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,2048,0.013251555462678274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,1024,0.014614222778214348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,2048,0.011351111034552256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,16384,0.04124800033039517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,64,0.0028008888992998335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,7168,0.21073333422342935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,65536,0.07636711332533094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,5120,0.019567999574873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,7168,0.010759111079904767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,2560,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,1024,0.004875555634498596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,65536,0.09967466857698228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,12288,0.01788444485929277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,256,0.003216000066863166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,4096,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,4096,0.02070488863521152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,12288,0.030046221282747056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,4096,0.01683288812637329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,512,0.004285333471165763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,3584,0.013774221969975365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,3584,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,3072,0.011746666497654386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,4096,0.03786311215824551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,65536,1.2353653377956815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,2560,0.012792000340090858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,64,0.02235200007756551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,4096,0.01700888905260298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,7168,0.011365333365069495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,3584,0.0377475553088718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,64,0.0028115556471877625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,7168,0.009708444277445475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,16384,0.03158844510714213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,3584,0.01147822207874722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,3072,0.008124444219801161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,1536,0.009165333376990424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,16384,0.04731555448638069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,12288,0.35950401094224715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,10240,0.07410399781333075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,5120,0.016892444756295946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,768,0.030011554559071858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,1024,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,12288,0.08694577879375881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,10240,0.016388444436921012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,7168,0.008656889200210571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,3584,0.019344889455371432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,64,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,512,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,4096,0.019321777754359774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,65536,13.429697672526041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,16384,0.03644266724586487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,512,0.01904800037542979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,128,0.0032151111712058387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,65536,0.30962755944993764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,2048,0.006924444602595435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,768,0.01666933298110962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,4096,0.011172444456153445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,3072,0.007669332954618666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,128,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,16384,0.029402666621738013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,5120,0.012809777425395118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,5120,0.01091466678513421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,3584,0.015576000014940897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,16384,0.021846221552954778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,512,0.011720000041855706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,3072,0.019560888409614563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,65536,0.028947555356555518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,2048,0.009315555294354757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,2560,0.12364266978369819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,1536,0.011191111471917896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,5120,0.015903110305468243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,64,0.010718222293588849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,1536,0.009944000177913243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,768,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,256,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,2048,0.007672888537247975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,768,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,512,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,768,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,3072,0.030455999904208716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,768,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,2560,0.013168888787428537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,7168,0.043881777260038585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,2560,0.02718222141265869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,3584,0.31857421663072377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,8192,0.014058666096793281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,512,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,2048,0.014104889498816596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,4096,0.011754666765530905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,5120,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,7168,0.03371999992264642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,10240,0.038093331787321306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,16384,0.021432888176706102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,3584,0.008628444539176093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,16384,0.01682933337158627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,64,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,12288,0.11609333091311985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,3584,0.03532622257868449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,16384,0.043976889716254346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,768,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,2560,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,256,0.015210666590266757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,64,0.005349333501524395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,2560,0.016515556308958266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,32,0.0034684443639384378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,2048,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,5120,0.019519999623298645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,128,0.0031786666562159858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,8192,0.022328888376553852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,6144,0.020795555578337777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,2048,0.006609777609507243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,768,0.017376888129446242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,5120,0.008393777741326226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,12288,0.24699910481770834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,3584,0.01684177749686771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,3072,0.06871466504202949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,128,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,4096,0.013101333545313941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,65536,0.059246222178141274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,10240,0.028044445647133723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,2048,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,10240,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,65536,0.22016710705227324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,7168,0.009992000129487779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,8192,0.023341332872708637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,128,0.0053226666318045715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,16384,0.019792889555295307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,8192,0.19228977627224395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,256,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,32,0.003527111063400904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,6144,0.025408888856569927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,12288,0.04393600092993843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,64,0.002667555585503578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,5120,0.10480799939897324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,3584,0.014522666732470194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,7168,0.016935999194780987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,3072,0.023433778021070693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,2048,0.015783111254374187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,128,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,2560,0.027599111199378967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,128,0.1693413390053643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,12288,0.03338399860594008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,10240,0.03085422184732225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,6144,0.01386222243309021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,256,0.006903999795516332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,32,0.007606222397751278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,1536,0.01056977775361803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,5120,0.021240888370407954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,128,0.012851555314328937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,256,0.005247999810510212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,512,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,256,0.013105777402718862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,16384,0.02069511181778378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,4096,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,3584,0.10429155826568604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,3584,0.029804444975323145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,7168,0.010062221851613786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,1536,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,5120,0.016723554995324876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,10240,0.018976888722843595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,10240,0.1560631063249376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,256,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,1024,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,256,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,10240,0.017314665847354464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,1536,0.014498665928840637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,3072,0.007275555696752336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,2560,0.007309333317809635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,16384,0.4170488781399197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,12288,0.031856000423431396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,768,0.0069004446268081665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,3072,0.04433866673045688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,768,0.0042275554604000514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,65536,0.13526933723025852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,512,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,4096,0.04517155554559496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,2048,0.006933333145247565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,8192,0.04216088851292928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,2048,0.0127368892232577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,2048,0.013436444103717804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,65536,2.934255176120334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,768,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,2048,0.016237333416938782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,3072,0.017112000121010672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,128,0.01513155632548862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,6144,0.008982222113344405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,65536,0.05752888653013441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,2560,0.03426577647527059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,6144,0.011390222443474663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,65536,0.06667733192443848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,1024,0.004282666577233208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,8192,0.034902221626705594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,32,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,5120,0.013803555733627744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,2560,0.017283555534150865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,512,0.00552977787123786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,2560,0.030943999687830608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,3072,0.015356444650226168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,3072,0.007993777592976889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,4096,0.1400808890660604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,1536,0.022781333989567224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,512,0.216575993431939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,8192,0.011410666836632622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,65536,0.4100622336069743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,7168,0.01180266671710544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,3072,0.036504000425338745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,1536,0.014447111222479077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,64,0.003536000019974179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,3584,0.007320889168315464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,12288,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,65536,0.0842497746149699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,8192,0.05365777677959866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,16384,0.013260444833172692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,3584,0.013912889692518445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,1024,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,2048,0.0158942225906584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,6144,0.01719377769364251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,2560,0.010477333433098264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,512,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,10240,0.027481777800454035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,256,0.005640888793600931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,3072,0.025711999999152288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,32,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,16384,0.018927999668651156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,10240,0.08857422404819065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,3584,0.11685510476430257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,2048,0.027288888891537983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,5120,0.007409777906205919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,1024,0.020303999384244282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,3584,0.008656000097592672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,256,0.005986666513813867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,3584,0.07539733250935872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,2560,0.06342933575312297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,16384,0.02700177828470866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,2560,0.007028444773621029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,3072,0.008797333472304875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,1536,0.01402933398882548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,5120,0.07095288568072848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,10240,0.01204355557759603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,65536,0.48376623789469403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,128,0.006911110960774952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,3584,0.009015999734401703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,7168,0.015171556009186639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,3584,0.008625778059164682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,2560,0.009306666751702627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,6144,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,512,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,12288,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,10240,0.025362667110231187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,65536,0.03363733490308126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,5120,0.016151999433835346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,10240,0.08076266447703044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,4096,0.0075235557225015424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,7168,0.015857777661747403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,6144,0.018762666318151686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,2048,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,6144,0.025754666990704004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,128,0.015202666322390238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,12288,0.013447999954223633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,2048,0.013106666505336761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,512,0.02444177700413598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,65536,0.8273359934488932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,5120,0.026241777671708003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,65536,0.3794257905748155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,5120,0.04527555571662056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,1536,0.04296711087226868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,65536,0.38371555010477704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,6144,0.18056089348263213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,256,0.005520888914664586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,5120,0.0188719994492001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,1536,0.016714667280515034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,2048,0.1267075538635254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,1024,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,7168,0.020230222079488967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,3072,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,3072,0.008663111262851292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,1536,0.17148533132341173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,5120,0.03596533338228861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,2560,0.02316177719169193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,8192,0.02071288890308804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,2560,0.015843555331230164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,5120,0.07975733280181885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,2560,0.021611554755104914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,4096,0.022286223040686712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,256,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,6144,0.023598222268952265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,10240,0.02922755479812622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,12288,0.014211555322011312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,12288,0.013431111143694984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,128,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,6144,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,128,0.005231110999981563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,3072,0.02815022071202596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,512,0.005423999908897612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,3072,0.08611910872989231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,10240,0.04435999857054817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,512,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,4096,0.013083555632167392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,8192,0.023002665903833177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,3584,0.013173333472675748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,128,0.003141333245568805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,16384,0.03307200140423245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,2048,0.016151999433835346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,12288,0.010285333626800114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,5120,0.06565244330300225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,64,0.0028435554769304064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,10240,0.09617066383361816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,6144,0.37850220998128253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,7168,0.017582222819328308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,10240,0.20145866605970594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,3072,0.016539555456903245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,3584,0.012416000167528788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,256,0.01613066593805949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,8192,0.09182844559351604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,1536,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,5120,0.025399110383457605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,7168,0.040291554398006864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,7168,0.019016888406541612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,8192,0.01072799993885888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,8192,0.03601333167817857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,2048,0.018752889500723947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,65536,0.08448000086678399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,8192,0.008006222546100616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,6144,0.014128888646761576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,3072,0.011869333684444427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,768,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,10240,0.014335110783576965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,3584,0.015820443630218506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,6144,0.08051111300786336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,2048,0.0074755557709270054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,3072,0.008348444269763099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,4096,0.034916443957222834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,64,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,768,0.05515733361244202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,5120,0.021500445074505273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,10240,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,32,0.003823111040724648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,512,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,768,0.007994666695594788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,8192,0.019263111882739596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,256,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,3584,0.007954667011896769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,1536,0.03369155526161194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,12288,0.04005600015322367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,256,0.00546488869521353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,4096,0.01648977730009291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,128,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,65536,0.6047413084242079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,12288,0.0284204449918535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,2560,0.1099626686837938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,10240,0.014495111174053617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,7168,0.08675022257698907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,2048,0.07008977731068929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,7168,0.007913777397738563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,4096,0.01787911189926995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,8192,0.03659822212325202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,16384,0.23222488827175566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,65536,0.014449778530332776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,768,0.01790488925245073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,3072,0.0086666668454806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,32,0.0028213332924577924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,768,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,12288,0.08204533656438191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,1024,0.015468445089128284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,3584,0.007637333538797166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,10240,0.023015111684799194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,1536,0.017229333519935608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,3072,0.11877777841356064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,2560,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,2048,0.00721066693464915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,128,0.005879111174080107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,7168,0.02629866699377696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,2560,0.010440888504187265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,1024,0.017860444055663217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,10240,0.01513155632548862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,5120,0.009870222045315636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,12288,0.014044443766276041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,10240,0.012057777908113269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,8192,0.03478399912516276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,4096,0.023471110396915015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,64,0.002806222273243798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,3584,0.008281778130266402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,4096,0.01574488812022739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,3072,0.016207110550668504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,64,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,10240,0.016659556163681876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,3584,0.023056000471115112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,1024,0.0038888889054457345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,12288,0.01862133377128177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,256,0.0028213332924577924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,6144,0.026540444956885442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,16384,0.01887911061445872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,3584,0.011000889043013254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,6144,0.026889777845806543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,4096,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,5120,0.008325333396593729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,10240,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,6144,0.0978542235162523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,3584,0.01886311173439026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,16384,0.0691937804222107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,64,0.002692444456948174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,8192,0.009317333499590555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,3584,0.093922668033176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,2048,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,4096,0.007602666815121968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,128,0.005222222043408289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,32,0.01037599974208408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,6144,0.02520177761713664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,32,0.0028479999552170434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,65536,0.014366222752465142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,12288,0.01309333327743742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,32,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,32,0.0032151111712058387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,5120,0.010513777534166971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,768,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,2048,0.013162666724787818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,65536,0.08605066935221355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,128,0.00833511104186376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,1024,0.02260977857642704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,1024,0.009323555562231276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,6144,0.021012443635198805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,256,0.023915555742051866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,6144,0.01310755560795466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,3584,0.022600000103314716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,4096,0.0070293330483966405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,7168,0.00941066692272822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,1536,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,32,0.017870222528775532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,1536,0.008635555704434713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,512,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,12288,0.019739554988013375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,256,0.0069466663731469055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,5120,0.012815111213260226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,12288,0.01757333344883389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,2560,0.02625244524743822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,2048,0.02013511127895779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,4096,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,12288,0.017864889568752713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,16384,0.10677866803275214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,3584,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,16384,0.016144000821643405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,16384,0.036461333433787026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,12288,0.027432888746261597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,32,0.002842666581273079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,3584,0.041689776711993747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,256,0.007634667058785756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,768,0.010400888820489248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,8192,0.013986666997273764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,5120,2.5349271562364364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,3072,0.007304888632562425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,768,0.03695911169052124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,8192,0.012413333687517377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,128,0.004848000076082018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,12288,0.26123732990688747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,512,0.01275111072593265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,12288,0.30535022417704266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,256,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,7168,0.010339555641015371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,16384,0.016788444585270353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,5120,0.01748533381356133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,768,0.008976000050703684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,3072,0.015078221758206686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,768,0.004581333448489507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,128,0.011352889239788055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,65536,0.06573955880271064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,768,0.014925332532988654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,2048,0.03384088807635837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,7168,0.010696888797812991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,4096,0.012619555824332766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,2560,0.018279999494552612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,12288,0.03270311156908671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,10240,0.05044088761011759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,10240,0.7272471321953667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,6144,0.039510223600599505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,3072,0.012426666915416718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,12288,0.013734222286277346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,64,0.010008888940016428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,512,0.00999466660949919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,512,0.008721777962313758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,768,0.01960355540116628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,64,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,128,0.014840000205569796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,8192,0.05900444587071737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,65536,0.023715555667877197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,1024,0.006679111056857639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,1536,0.018565333551830716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,5120,0.048073778549830117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,8192,0.00833333366447025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,10240,0.011792889071835412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,5120,0.017088888419999015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,7168,0.03964266512129042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,1024,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,1536,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,3584,0.015525332755512662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,3584,0.02289955483542548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,8192,0.06586933135986328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,5120,0.23991823196411133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,7168,0.009950222240553962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,16384,0.048785779211256236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,6144,0.008975110948085785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,4096,0.008980444735950893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,2048,0.005196444276306364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,7168,0.017542221479945712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,8192,0.01202311118443807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,3584,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,16384,0.035417778624428645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,1536,0.007138667007287343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,128,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,768,0.04424711068471273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,5120,0.011920889218648275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,65536,0.19817422495947945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,8192,0.06785955693986681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,256,0.008000888758235509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,2560,0.013407111167907715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,3584,0.027798222170935735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,16384,0.010871999793582492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,16384,0.05762133333418104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,128,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,3072,0.02219466699494256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,2560,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,8192,0.05750844213697645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,32,0.0031973332580592898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,12288,0.01921422282854716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,2560,0.013154666456911298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,7168,0.07168000274234347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,8192,0.010337777435779572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,6144,0.007882666256692674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,768,0.01681422193845113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,512,0.005541333307822545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,10240,0.20489511224958631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,6144,0.04163644380039639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,2560,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,8192,0.018903111418088276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,10240,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,3072,0.0729279981719123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,16384,0.04617955618434482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,1536,0.011386666860845355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,3584,0.015847999188635085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,256,0.01755555636352963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,2560,0.016153777639071148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,7168,0.09558844566345215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,1536,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,2048,0.012383999923865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,256,0.00565155554148886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,1536,0.09963199827406143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,3584,0.039804445372687444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,8192,0.023941333095232647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,16384,0.011014222270912595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,1536,0.02745955520206028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,1536,0.08909333414501613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,4096,0.02033866610791948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,64,0.008098666866620382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,2048,0.008882666627566019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,4096,0.009352000223265754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,3072,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,1024,0.006250666661394968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,12288,0.028011557128694322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,5120,0.02277244461907281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,512,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,65536,0.07515733109580146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,256,0.004403555558787452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,65536,0.09973066382937962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,4096,0.008644444247086843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,5120,0.008386666576067606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,5120,0.017820444371965196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,6144,0.06741244263119169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,10240,0.055227554506725736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,1024,0.014843554960356818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,3072,0.012443555725945367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,128,0.0031226666437255014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,10240,0.9425955878363715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,6144,0.342865784962972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,256,0.016156444946924847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,1024,0.01613333324591319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,8192,0.17645155058966744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,32,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,32,0.0027439999911520216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,2048,0.0052879999081293745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,8192,0.010382222632567087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,256,0.017165333032608032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,6144,0.012527999778588613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,65536,0.06755022207895915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,3072,0.007145778172545963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,3072,0.015948444604873657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,7168,0.010269333091047075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,1536,0.008097777764002482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,12288,0.13106844160291883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,2048,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,10240,0.02602844436963399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,512,0.0032382222513357797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,1536,0.013835555149449242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,65536,0.2902719974517822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,1024,0.015824889143308003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,3072,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,3584,0.008967111508051554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,2048,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,64,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,16384,0.07504177755779691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,2560,0.009008000294367472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,3584,0.009959110783206092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,8192,0.022611555125978258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,16384,0.03029777937465244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,768,0.0042035554846127825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,768,0.0069120000633928515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,12288,0.1851137744055854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,16384,0.012999110751681857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,768,0.016535111599498324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,2560,0.007311999797821045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,32,0.0038924444880750445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,1536,0.01443555619981554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,2560,0.01791288952032725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,256,0.005636444522274866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,12288,0.019340443942281935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,10240,0.020581333173645865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,16384,0.02675466736157735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,1024,0.006522666662931442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,16384,0.015433778365453085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,3072,0.027223111854659185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,10240,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,128,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,10240,0.05320622192488777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,2048,0.006709333509206772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,4096,0.01905866629547543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,10240,0.018686221705542672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,256,0.02195377813445197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,32,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,2560,0.008614222208658854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,2048,0.00685066646999783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,512,0.017384888397322763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,65536,0.1336711115307278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,16384,0.015251555376582675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,7168,0.3746284378899469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,2048,0.008737777670224508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,12288,0.029185778564876978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,6144,0.009660444325870937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,128,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,2048,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,7168,0.0306915541489919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,512,0.026745777991082933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,2560,0.015488000379668342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,5120,0.058033777607811816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,32,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,3072,0.022081777453422546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,2048,0.02217777735657162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,8192,0.018017777138286166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,768,0.004577777865860197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,1024,0.012120000190205045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,32,0.009334222310119206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,2048,0.02831288841035631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,6144,0.9709901809692383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,8192,0.3455395433637831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,4096,0.01551644504070282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,512,0.015171556009186639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,10240,0.03675644596417745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,512,0.016840888394249808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,128,0.013551111022631327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,4096,0.019727999965349834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,256,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,2048,0.014760888285107084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,2048,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,1024,0.00924177798959944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,4096,0.008763555851247575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,4096,0.011795555551846823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,5120,0.04814577764934964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,5120,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,2560,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,3072,0.05172888769043816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,3072,0.007671111159854465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,6144,0.025240888198216755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,65536,0.0947306685977512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,128,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,128,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,4096,0.008662222160233391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,512,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,6144,0.019807111885812547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,8192,0.03699466586112976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,6144,0.015078221758206686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,12288,0.0964311096403334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,10240,0.0363377763165368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,16384,0.03820533222622342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,4096,0.008019555773999956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,512,0.01852888862291972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,6144,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,128,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,8192,0.011048888994587792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,7168,0.012456888953844706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,2048,0.006986666884687211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,4096,0.00925511121749878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,768,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,65536,0.06284444199668036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,256,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,128,0.006471111128727595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,32,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,65536,0.09012977944480048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,256,0.020596444606781006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,12288,0.029876444074842665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,256,0.012821333275900947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,128,0.028395556741290625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,7168,0.026914666096369427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,6144,0.05460177858670553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,6144,0.015788444214397006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,10240,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,1536,0.004521777646409141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,6144,0.04203377829657661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,512,0.010152888794740042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,512,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,8192,0.01479822231663598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,3072,0.007273777491516537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,12288,0.1559644407696194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,5120,0.13048356109195286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,3584,0.007617777420414819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,768,0.005716444303592046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,2560,0.05122933453983731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,65536,0.4710542360941569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,256,0.00351200004418691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,768,0.01906222270594703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,65536,0.18458844555748832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,4096,0.017126222451527912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,3072,0.06150044335259331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,1536,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,6144,0.025802666942278545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,256,0.005561777700980504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,12288,0.013108444710572561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,12288,0.2296595573425293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,6144,0.5307617717319065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,64,0.0032346666687064697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,4096,0.0657857788933648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,7168,0.01016355554262797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,1536,0.013028444515334235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,1024,0.006960888703664144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,3072,0.0074551113777690465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,8192,0.042605333858066134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,1024,0.0307608875963423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,512,0.015096889601813423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,32,0.0031848889258172777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,65536,0.2167511118782891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,768,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,7168,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,128,0.003572444534964032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,256,0.0042275554604000514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,7168,0.032872888776991106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,8192,0.03664355476697286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,12288,0.03559022148450216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,16384,0.023473777704768713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,2560,0.025986666480700176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,6144,0.012406222522258759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,256,0.02508977717823452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,512,0.014198222094111972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,4096,0.013474666410022311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,5120,0.029973334736294214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,256,0.02882844540807936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,5120,0.014719999498791166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,3072,0.011050666371981302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,256,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,128,0.006569777925809224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,32,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,65536,0.5367502106560601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,512,0.006108444597986009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,3584,0.010168000227875179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,5120,0.009359999663299983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,2048,0.006879111131032308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,3584,0.01391377713945177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,8192,0.023375110493765935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,256,0.015098666151364645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,8192,0.024346666203604803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,2048,0.0073315559162033936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,2048,0.009873777627944946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,8192,0.03966844413015578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,128,0.003569777641031477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,8192,0.05688533518049452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,12288,0.04962933394643995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,12288,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,1024,0.005992888990375731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,2560,0.008292444050312042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,16384,0.027493332823117573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,1024,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,5120,0.009401777552233802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,2048,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,5120,0.012705778082211813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,1024,0.005768888940413793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,256,0.017567111386193168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,65536,0.15740977393256292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,1536,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,2048,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,12288,0.01796266602145301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,10240,0.020976889464590285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,4096,0.025055110454559326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,512,0.005532444351249271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,256,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,3072,0.021339555581410725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,512,0.009647111097971598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,256,0.0057653333577844836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,65536,0.07194488578372531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,65536,0.06595289044910006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,768,0.01126577787929111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,1024,0.007268444531493717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,12288,0.028010666370391846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,5120,0.024265777733590867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,1024,0.00423911131090588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,1536,0.004560889055331548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,768,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,128,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,10240,0.047953777843051486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,256,0.012065778175989786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,1024,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,16384,0.03655644589000278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,16384,0.2644791073269314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,256,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,3584,0.015815110670195687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,512,0.004916444420814514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,12288,0.014074666632546319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,8192,0.03589866558710734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,2560,0.007273777491516537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,1536,0.004515555583768421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,1536,0.004398222184843487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,128,0.004294222427739037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,4096,0.06919555531607734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,6144,0.01754666699303521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,8192,0.016770665844281513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,5120,0.009008888569143083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,32,0.003128888913326793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,128,0.00535733335547977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,8192,0.016526222229003906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,1024,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,768,0.005311111195219888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,512,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,1536,0.004538666870858935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,10240,0.14675288730197483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,4096,0.03633600142267015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,64,0.006592000110281839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,2560,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,16384,0.02440799938307868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,64,0.005701333284378052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,2560,0.007911110917727152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,512,0.004859555512666702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,6144,0.0673875543806288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,512,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,5120,0.015593777100245157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,1536,0.004576888763242298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,768,0.014486221803559197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,256,0.06017511420779758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,7168,0.01925066610177358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,64,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,10240,0.08601333035363092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,32,0.007272889216740926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,768,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,256,0.00349866681628757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,2048,0.024846222665574815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,3072,0.008921778036488427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,5120,0.01741333305835724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,12288,0.028762665059831407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,64,0.0025191110455327565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,16384,0.11281243960062663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,256,0.006703111032644908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,6144,0.016199111938476562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,768,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,1536,0.035190221336152815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,3584,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,4096,0.007968000239796108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,256,0.012151999606026543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,64,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,16384,0.1678666671117147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,65536,0.21481777562035453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,2048,0.029529776838090684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,12288,0.058648890919155545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,32,0.003177777760558658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,4096,0.24769777721828887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,7168,0.022816000713242426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,128,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,4096,0.11246311664581299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,6144,0.011447110937701331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,8192,0.023080888721677992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,32,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,64,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,7168,0.016968000266287062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,256,0.0034826666944556763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,3072,0.03401155604256524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,3584,0.014753777119848462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,6144,0.011576889289749993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,2048,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,8192,0.02362577782736884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,64,0.005527110977305307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,6144,0.01889333294497596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,12288,0.014160000615649752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,256,0.0028773333049482773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,8192,0.11773688263363308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,4096,0.029508444997999404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,8192,0.03461777832772996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,12288,0.027836445305082533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,3584,0.046560890144772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,16384,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,7168,0.011063111325105032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,65536,0.010755555497275459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,3584,0.007335111498832703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,768,0.026854222019513447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,3584,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,2560,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,512,0.03552622265285916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,2048,0.04865600003136528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,256,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,256,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,6144,0.02072355482313368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,5120,0.0814995567003886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,5120,0.03334488802485996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,7168,0.0188773340649075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,2048,0.014211555322011312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,12288,0.024359999431504145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,8192,0.0123848890264829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,2048,0.013095999757448832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,768,0.01723288827472263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,5120,0.041011555327309504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,512,0.00795199970404307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,16384,0.021330666210916307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,65536,0.6834835476345487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,1536,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,1024,0.021690666675567627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,512,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,3072,0.014334221680959066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,10240,0.02614044480853611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,768,0.0041742221348815495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,10240,0.0830942193667094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,128,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,32,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,12288,0.033246222469541765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,2048,0.013047110703256396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,12288,0.14673599931928846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,32,0.007974222302436829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,2048,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,512,0.006044444524579578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,5120,0.016553777787420485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,5120,0.024600888291994732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,12288,0.04368266794416639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,7168,0.013493333425786761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,6144,0.009034666750166152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,1536,0.01925244430700938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,2560,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,2560,0.06198133362664116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,64,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,16384,0.021168000168270532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,3072,0.03447555502255758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,512,0.021367111139827307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,4096,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,8192,0.01351377781894472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,64,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,65536,0.08289155695173475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,256,0.009798222117953831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,512,0.03161688976817661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,256,0.0037742224004533556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,5120,0.054901334974500865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,8192,0.01586666703224182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,64,0.006636444479227066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,16384,0.011735111474990845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,12288,0.027842667367723253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,3584,0.008263111114501953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,10240,0.04389866524272495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,3072,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,16384,0.03227644496493869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,512,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,16384,0.02162666618824005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,16384,0.08702488740285237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,10240,0.013809777796268463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,3584,0.011747555600272285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,4096,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,1024,0.009411555197503831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,32,0.002846222163902389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,8192,0.03362755642996894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,64,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,3072,0.06459466616312663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,768,0.030055999755859375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,7168,0.011418666276666852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,8192,0.013077333569526672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,8192,0.021990221407678392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,7168,0.014533332652515836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,256,0.015178667174445258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,7168,0.025311999850802954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,512,0.0058755555914507965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,2048,0.014374223020341663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,7168,0.0218231115076277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,2560,0.013850666582584381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,10240,0.026037333740128413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,3072,0.011904000408119626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,512,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,3584,0.015800888339678448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,1024,0.0076453329788313965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,1024,0.0038480001191298165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,3072,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,5120,0.009754666851626502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,512,0.006024888820118374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,128,0.011400889191362592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,1024,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,3584,0.01682933337158627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,32,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,128,0.009775111244784461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,512,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,256,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,768,0.017145777742067974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,3584,0.0075697774688402815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,16384,0.04580977890226576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,5120,0.012855999999576144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,128,0.014516444669829475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,768,0.008311111066076491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,65536,0.1847840017742581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,65536,0.07030400302675036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,64,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,5120,0.028021332290437486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,10240,0.016844444804721408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,10240,0.12720621956719294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,1536,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,64,0.0028799999919202593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,3072,0.01551999979548984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,1536,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,1536,0.014889778362380134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,1536,0.016384000579516094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,12288,0.026742221580611333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,512,0.005215999980767568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,768,0.016535111599498324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,7168,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,65536,0.1946106619305081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,32,0.024754666619830664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,8192,0.38433067003885907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,256,0.003530666646030214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,12288,0.012098666694429187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,10240,0.16035021675957575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,32,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,1536,0.006758222149478064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,3584,0.007747555772463481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,8192,0.6159751150343153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,128,0.061421334743499756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,1536,0.01510933372709486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,2048,0.011024889018800525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,16384,0.016161777906947665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,2560,0.06364177995257907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,2560,0.016460445192125108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,10240,0.01759199963675605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,5120,0.16025155120425755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,4096,0.01852888862291972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,256,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,12288,0.06616533464855619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,7168,0.07829866806666057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,10240,0.03203466534614563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,3584,0.02604177759753333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,2560,0.010195555786291758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,6144,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,12288,0.018624888526068795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,8192,0.012710221939616732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,2560,0.02269688910908169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,65536,0.1389875544442071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,1536,0.006341333190600078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,65536,2.563892364501953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,3072,0.01036444471942054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,32,0.003855111284388436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,64,0.017533333765135873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,2048,0.02310666607485877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,5120,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,2048,0.015472888946533203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,128,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,1024,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,3072,0.017335111896197002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,64,0.0028000000036425064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,7168,0.013071999781661563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,16384,0.07544266515307956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,128,0.008606221940782335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,1536,0.007261333366235097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,2048,0.023426666855812073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,7168,0.14432888560824925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,3584,0.13080710834927028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,2560,0.00979288915793101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,4096,0.01737955543729994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,65536,0.11652978261311848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,2560,0.013280889226330651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,6144,0.009631111390060848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,1024,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,12288,0.23567732175191244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,32,0.00301155551440186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,256,0.016192000773217943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,7168,0.024706666668256123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,2560,0.007101332975758447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,768,0.007251555720965068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,4096,0.018975999620225694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,3072,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,65536,0.07928444279564752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,3584,0.02014488809638553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,3072,0.016018667154841952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,3584,0.03397511111365424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,12288,0.17993777328067354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,128,0.003510222252872255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,5120,0.038487109873029925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,2560,0.009306666751702627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,32,0.0028284444577164124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,1536,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,10240,0.01919822229279412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,8192,0.013096888860066732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,12288,0.01770400007565816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,5120,0.008616000413894653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,4096,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,65536,0.08955999877717759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,1536,0.028168890211317275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,32,0.006615110983451207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,2048,0.01219377749496036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,3072,0.2949377695719401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,256,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,32,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,8192,0.07625689109166463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,2048,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,65536,0.06713777780532837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,32,0.0028213332924577924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,128,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,5120,0.008433777425024245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,3584,0.015116444892353482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,128,0.01443555619981554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,5120,0.01128888875246048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,6144,0.0697751111454434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,1024,0.00924622184700436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,7168,0.023375999596383836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,64,0.0024986666523747975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,32,0.0034773333205117118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,2048,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,32,0.011015999648306104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,512,0.0373342235883077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,6144,0.01966399947802226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,768,0.013144888811641268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,2560,0.5210009151034886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,12288,0.066520889600118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,4096,0.011733333269755045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,768,0.03001777662171258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,4096,0.00872444444232517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,8192,0.01629866659641266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,768,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,128,0.0031395554542541504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,2048,0.045695109499825366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,5120,0.009693333672152625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,32,0.0063866666621632045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,1536,0.030851556195153132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,768,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,12288,0.03416444526778327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,2560,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,2560,0.013781333135233985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,3584,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,768,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,12288,0.02202044427394867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,256,0.009655111365848118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,12288,0.02398577829202016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,3072,0.01752444439464145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,128,0.008283555507659912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,16384,1.1398702197604709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,7168,0.021772444248199463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,1024,0.005864000154866113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,128,0.007999999655617608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,1024,0.006009777800904379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,256,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,7168,0.009711999860074785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,4096,0.016207110550668504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,768,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,256,0.003156444471743372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,1024,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,10240,0.08022666639751859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,10240,1.1104577382405598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,2048,0.012540444731712341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,6144,0.01699911057949066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,2048,0.10035200251473321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,2048,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,1024,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,128,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,5120,0.02477600011560652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,768,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,12288,0.036807109912236534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,6144,0.022632888621754114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,1024,0.028168890211317275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,512,0.0273635552989112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,2048,0.009000889129108852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,5120,0.15198311540815565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,6144,0.009157333109113906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,512,0.003556444413132138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,7168,0.2990960015190972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,1536,0.01585422290696038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,1024,0.00572622236278322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,7168,0.038839999172422625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,6144,0.01279022213485506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,16384,0.08556266625722249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,1536,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,7168,0.010077333284748925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,1024,0.011576000187132092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,16384,0.03156444430351257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,3072,0.007311999797821045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,65536,0.19396444161732992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,1536,0.006395555618736479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,1536,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,3584,0.018263111511866253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,768,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,5120,0.024634665913052026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,5120,0.03091555502679613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,8192,0.01035733355416192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,1536,0.0064106666379504735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,2560,0.03629866573545668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,768,0.03035822179582384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,4096,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,32,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,65536,0.044904887676239014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,1536,0.022841778066423204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,3072,0.014194667339324951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,64,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,16384,0.045427554183536105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,7168,0.010006222460005019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,128,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,1536,0.0046453335218959385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,7168,0.012282667060693106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,32,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,12288,0.13250933753119573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,32,0.002936888900068071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,12288,0.0606293347146776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,32,0.00795288880666097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,128,0.004224888980388641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,2560,0.006672888994216919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,2560,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,4096,0.015757333901193406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,7168,0.179730667008294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,128,0.03652888867590163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,768,0.01793511046303643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,256,0.023015111684799194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,512,0.0031866667171319327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,128,0.011121778024567498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,3584,0.1689902279112074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,768,0.04140977727042304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,65536,0.31167909834120006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,2560,0.017536888519922893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,128,0.030527111556794908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,10240,0.017989334132936265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,65536,0.15758400493197972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,16384,0.021500445074505273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,12288,0.014074666632546319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,10240,0.011732444167137146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,8192,0.05024888780381945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,12288,0.03733333283000522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,1536,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,8192,0.017286222842004564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,1024,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,65536,0.14821599589453802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,2048,0.01314222233162986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,3072,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,128,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,2048,0.024400888217820063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,128,0.014834667245546976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,65536,0.11896977159712051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,512,0.013516444298956128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,3584,0.008151111503442129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,8192,0.016255110502243042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,4096,0.008430221842394935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,12288,0.01722577710946401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,768,0.014326221413082547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,3584,0.016957332690556843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,8192,0.010681777364677854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,1024,0.004567999806669024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,128,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,128,0.01276622215906779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,512,0.01348088930050532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,256,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,8192,0.016786666380034555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,256,0.006729777902364731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,256,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,65536,0.04541866646872627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,16384,0.015847999188635085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,32,0.01903555625014835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,1536,0.013787555197874704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,32,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,65536,0.16726311047871908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,12288,0.040588445133633085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,12288,0.4021608829498291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,6144,0.028062221076753404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,5120,0.016872000363137987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,2560,0.015270221564504834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,1536,0.006250666661394968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,6144,0.044473777214686074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,768,0.013642667068375481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,16384,0.031543999910354614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,512,0.012411555482281579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,5120,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,1536,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,256,0.020273778173658583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,65536,0.0637413329548306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,8192,0.010412444670995077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,128,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,10240,0.012040889097584618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,2048,0.011788444386588203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,8192,0.029296000798543293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,8192,0.01484444406297472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,65536,0.586334228515625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,7168,0.06588088803821139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,768,0.018664000762833487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,3584,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,768,0.009001777403884465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,65536,0.08177244663238525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,1024,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,256,0.0032755554550223877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,128,0.0031555555760860443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,512,0.014468444718254937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,5120,0.007634667058785756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,65536,0.06428355640835233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,4096,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,4096,0.008435555630260045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,16384,0.04985510971811083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,2560,0.009735110733244155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,6144,0.008357333640257517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,65536,0.11531733142005073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,7168,0.09595733218722874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,768,0.005684444473849402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,2048,0.017589333984586928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,7168,0.010013333625263637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,3072,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,7168,0.02679377794265747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,1536,0.0052915554907586836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,4096,0.007667555577225155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,65536,0.3240399890475803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,2560,0.015480000111791821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,16384,0.03888355692227682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,65536,0.08265155553817749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,2560,0.009383999639087254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,16384,0.49818934334648984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,5120,0.010844444235165914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,12288,0.01514755520555708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,1536,0.03798933492766486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,4096,0.030647112263573542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,8192,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,1536,0.16921777195400664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,65536,0.3450506793128119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,6144,0.04474488894144694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,16384,0.024329778220918443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,2560,0.007350222104125553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,2048,0.024316444993019104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,2048,0.005039999882380168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,5120,0.11174578136867946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,10240,0.025293333662880793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,4096,0.06919288635253906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,32,0.0028986665937635633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,16384,0.016176889340082806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,7168,0.07436266872617933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,7168,0.009474666582213508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,5120,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,64,0.011045333411958484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,5120,0.010074666804737514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,3584,0.005993777679072485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,256,0.016903110676341586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,4096,0.01440355512830946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,3584,0.02362488872475094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,7168,0.07080800003475614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,3584,0.014648000399271647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,65536,0.015767999821239047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,1024,0.009344889058007134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,16384,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,128,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,4096,0.014530667000346713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,4096,0.01796888808409373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,65536,0.4526009029812283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,65536,0.07220621903737386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,32,0.0037128888070583344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,768,0.01484266585773892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,64,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,6144,0.013212444053755866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,1024,0.018200889229774475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,16384,0.03023911184734768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,2048,0.009815110928482479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,128,0.0052879999081293745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,512,0.004263999975389904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,512,0.008360000120268928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,2048,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,2048,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,768,0.007191999918884701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,2048,0.021465778350830078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,64,0.0028737777223189673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,512,0.005649777750174205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,256,0.10648622115453084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,1024,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,512,0.014843554960356818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,256,0.003941333128346337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,128,0.010479110810491772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,1536,0.012455999851226807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,128,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,1536,0.0255022214518653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,1536,0.006406222366624408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,65536,0.535261313120524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,16384,0.02982399861017863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,12288,0.2684195571475559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,65536,2.9042710198296438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,2048,0.011129777464601727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,8192,0.04164177841610379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,2048,0.018588445252842374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,768,0.012823999755912356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,7168,0.05308711197641161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,2560,0.04274400075276693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,8192,0.09343733390172322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,6144,0.019004444281260174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,12288,0.028000887897279527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,3072,0.018984888990720112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,7168,0.025260445144441392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,16384,0.2986977895100911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,4096,0.01519733336236742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,4096,0.007988444632954067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,512,0.006350222147173352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,128,0.0033297776761982176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,1024,0.00600000015563435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,2048,0.005062222066852781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,6144,0.05606933434804281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,16384,0.08161422279145983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,7168,0.024900444679790076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,2048,0.018252443936136033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,3072,0.025756445195939805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,12288,0.0399528874291314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,256,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,4096,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,2048,0.07043111324310303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,3072,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,1536,0.0094355551732911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,2560,0.019181332654423185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,7168,0.019354666272799175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,128,0.003160888950030009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,128,0.005398222141795688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,65536,0.1308364470799764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,16384,0.038400888442993164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,256,0.01684177749686771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,2048,0.010272888673676385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,10240,0.012115555504957834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,10240,0.0826666686269972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,512,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,10240,0.01802399920092689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,1024,0.01926933394538032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,6144,0.13023200300004747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,768,0.005061333378156026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,5120,0.117850661277771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,3584,0.010648888846238455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,6144,0.011499555574523078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,4096,0.0322693321439955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,12288,0.03681333197487725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,2048,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,768,0.007716444631417592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,12288,0.016787555482652452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,16384,0.029896001021067303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,6144,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,6144,0.031855109665128924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,8192,0.03258311086230808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,2048,0.11868000030517578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,1024,0.006047111004590988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,768,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,768,0.009340444372759925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,8192,0.015415110521846347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,256,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,1536,0.018432888719770644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,3584,0.12614133622911242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,7168,0.015122666954994202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,10240,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,10240,0.030187557140986126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,1024,0.04561777909596761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,128,0.015469332536061605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,10240,0.012135110795497894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,3584,0.015587555037604438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,2048,0.015203555425008139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,128,0.0029128889242808023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,6144,0.016706667012638517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,3072,0.031120889716678198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,2560,0.0086666668454806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,7168,0.02305511136849721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,3072,0.00813066628244188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,8192,0.12580977545844182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,6144,0.02362844513522254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,128,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,1536,0.006709333509206772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,8192,0.011393778026103973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,7168,0.006719111154476802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,256,0.013974222871992322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,64,0.0028408887899584244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,1024,0.007376888559924231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,8192,0.0307279990779029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,3584,0.02367555598417918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,3072,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,8192,0.07168533404668172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,4096,0.010072888599501716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,12288,0.027307555079460144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,10240,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,6144,0.011115555961926779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,8192,0.014840889308187695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,8192,0.01997155613369412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,64,0.00554577757914861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,512,0.013833777772055732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,1536,0.014156444205178155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,2560,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,7168,0.01609777741962009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,5120,0.07856088876724243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,768,0.020089776979552377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,256,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,65536,0.05681155787573921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,3072,0.011008888483047485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,10240,0.02529777752028571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,3584,0.008715555899673039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,2048,0.006341333190600078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,3072,0.02739199995994568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,768,0.01960266629854838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,5120,0.017255999975734286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,6144,0.009378666679064432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,16384,0.02749866743882497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,12288,0.033336000310050115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,256,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,768,0.014843554960356818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,768,0.00684799998998642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,5120,0.018592889110247295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,5120,0.01426666643884447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,2560,0.013767110804716745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,12288,0.043862223625183105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,8192,0.011457777685589261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,1024,0.01683466633160909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,5120,0.016513778103722464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,16384,0.05008000135421753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,65536,0.12743200196160212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,2560,0.24290222591824004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,2048,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,65536,0.06240266561508179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,4096,0.0092275556590822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,10240,0.015778667396969266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,65536,0.1402719948026869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,10240,0.011780444118711682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,2048,0.01387466655837165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,6144,0.0203475554784139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,10240,0.5165351231892904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,7168,0.014510222607188754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,8192,0.023335110810067918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,1024,0.015471110741297403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,1536,0.015544000599119397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,7168,0.012873777912722694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,65536,0.07224888934029473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,5120,0.011332444846630096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,8192,0.04388533367051018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,3072,0.01036800030204985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,768,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,1024,0.00461777796347936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,2048,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,128,0.005241777747869492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,10240,0.030696001317765977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,256,0.010969777901967367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,7168,0.00980088859796524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,16384,0.12893066141340467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,4096,0.03882222374280294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,4096,0.10621333122253418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,1536,0.0042408886882993905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,5120,0.011411555939250521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,4096,0.01072177787621816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,4096,0.030603554513719346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,3584,0.02741688821050856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,2560,0.011086222198274402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,128,0.006934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,2560,0.03650488787227207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,1024,0.006311111152172089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,5120,0.08203822374343872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,256,0.00425155543618732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,7168,0.22656801011827257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,5120,0.01866222255759769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,6144,0.02147644427087572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,1536,0.012129777835475074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,10240,0.08231821987364027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,10240,0.03052355514632331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,64,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,6144,0.02607822252644433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,7168,0.010018666585286459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,2048,0.008001777860853408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,10240,0.031609776947233416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,1536,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,16384,0.016887111796273124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,2048,0.017872000734011333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,2048,0.006983111302057902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,128,0.01109866632355584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,4096,0.01314755529165268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,5120,0.016247999336984422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,128,0.029241777128643457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,12288,0.28005867534213597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,3584,0.017606221967273288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,5120,0.027201778358883325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,4096,0.01405066748460134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,1536,0.010457778142558204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,256,0.01551644504070282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,64,0.0028684443483750024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,8192,0.01347733371787601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,3072,0.007965332931942409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,3584,0.018292443619834054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,65536,0.4163795577155219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,8192,0.005626666463083691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,3584,0.01017599966790941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,128,0.0029022221763928733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,6144,0.017255110873116385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,65536,0.9036551581488715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,2560,0.0079644446571668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,10240,0.012489777472284106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,16384,0.0070088886552386815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,3072,0.0415013333161672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,2048,0.007349333001507654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,64,0.00563466673096021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,32,0.004649777793222004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,2048,0.007353777686754863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,8192,0.018954666124449838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,1024,0.007054222126801808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,8192,0.010747555229398938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,10240,0.036722666687435575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,8192,0.018783110711309645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,768,0.017512889371977914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,128,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,10240,0.025480000509156123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,10240,0.016200888488027785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,65536,0.05073066552480062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,2048,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,1536,0.06788977649476793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,2560,0.01717688971095615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,256,0.0032337777730491427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,7168,0.00979911122057173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,5120,0.008912888665994009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,1024,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,3584,0.019188443819681805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,12288,0.019633778267436557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,12288,0.018017777138286166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,2048,0.008346666892369589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,7168,0.02222488820552826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,7168,0.010054222411579555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,1024,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,16384,0.01685688893000285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,2560,0.005767999837795894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,4096,0.020596444606781006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,8192,0.022472888231277466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,6144,0.09565333525339763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,3072,0.021176889538764954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,1536,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,128,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,3072,0.028621332512961492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,6144,0.012216000093354119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,768,0.017190222938855488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,4096,0.026760000321600173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,16384,0.02161600026819441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,1024,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,7168,0.009783999787436591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,2560,0.006734222173690796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,12288,0.020561777883105807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,3072,0.007999999655617608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,2048,0.014530667000346713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,3072,0.013945778210957846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,3072,0.0070053330726093715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,3072,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,512,0.0069955554273393415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,6144,0.018935999936527677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,768,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,5120,0.1339199940363566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,10240,0.0746142201953464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,768,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,5120,0.14967555469936794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,8192,0.01108088923825158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,3584,0.011749332977665795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,3584,0.015542222393883599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,16384,0.09886755545934041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,64,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,256,0.005992888990375731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,32,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,4096,0.010423111418883005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,12288,0.2614942126803928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,7168,0.01371022231049008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,2048,0.005773333211739858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,512,0.015068444940778943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,64,0.003935111065705617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,512,0.007739555504586961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,16384,0.016823111308945548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,512,0.012125333150227865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,12288,0.013983110586802164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,5120,0.008527999950779809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,4096,0.008352888955010308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,4096,0.011145778000354767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,32,0.006239999913507038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,32,0.003531555748648114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,64,0.018741332822375827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,6144,0.009136888715955947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,10240,0.01996799972322252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,16384,0.02035466664367252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,6144,0.00902133352226681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,5120,0.008760000268618265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,16384,0.007340444458855524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,65536,0.1740515496995714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,16384,0.05109066764513651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,2560,0.007364444434642792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,2048,0.008126222425036961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,3072,0.019641778535313077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,256,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,1024,0.005646222167544895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,3584,0.019877332780096266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,1024,0.015544889701737298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,1024,0.004271999829345279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,3584,0.014835556348164877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,32,0.049248887432946094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,7168,0.03656533360481262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,16384,0.008985777695973715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,2560,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,65536,0.125690672132704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,3072,0.009880888793203566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,512,0.022304889228608873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,5120,0.015698666373888653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,12288,0.1254426638285319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,10240,0.01756977703836229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,768,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,7168,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,32,0.0029191111938820947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,1536,0.018253333038753934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,10240,0.016937777400016785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,1024,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,512,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,3072,0.017625777257813346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,16384,0.03780711028310988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,12288,0.019277332557572257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,7168,0.012799999780125089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,768,0.017197334104114108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,2560,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,1024,0.017626666360431247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,128,0.0032506665835777917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,3584,0.015633778439627755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,3072,0.00700800038046307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,65536,0.26838665538363987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,768,0.00564177789621883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,12288,0.02100177771515316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,3072,0.019874667127927143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,768,0.010505778094132742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,16384,0.4693022304111057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,2560,0.007719111111429002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,12288,0.025772444076008264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,10240,0.13264711697896323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,256,0.003535110917356279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,2560,0.02450399928622776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,2560,0.15050133069356283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,512,0.003913777983850903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,5120,0.012167111039161682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,3072,0.16003999445173475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,64,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,16384,0.014471999473041959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,8192,0.02882400155067444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,2048,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,2048,0.10272622108459473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,3584,0.007743111087216272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,16384,0.11657688352796768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,7168,0.013209777573744455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,16384,0.022892443670166865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,2048,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,8192,0.14004533820682102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,2048,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,512,0.013762666947311826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,2560,0.027503111296229895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,64,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,65536,0.07237066825230916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,2048,0.01143555591503779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,5120,0.08233244551552667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,1536,0.021320889393488567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,64,0.0031351111829280853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,1536,0.00664444433318244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,2560,0.015383111106024848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,512,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,65536,0.016559999850061204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,2048,0.008960888617568547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,2048,0.016558221644825406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,16384,0.01927199959754944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,64,0.005848000033034219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,2560,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,6144,0.0196088883611891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,64,0.015806222955385845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,4096,0.0321706665886773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,128,0.027816888358857896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,8192,0.027093332674768236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,3584,0.01332800007528729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,3584,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,256,0.015145777000321282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,512,0.007309333317809635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,16384,0.027740443746248882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,256,0.01566133399804433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,512,0.0032355555643637977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,65536,0.07601866457197401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,8192,0.032087110810809664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,128,0.005316444569163852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,4096,0.014175110393100314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,3584,0.0240631103515625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,128,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,12288,0.020945777495702107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,12288,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,10240,0.06593955225414701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,1024,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,3072,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,6144,0.01142133358452055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,65536,2.1832781897650824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,128,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,1024,0.006334222025341458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,768,0.003903111235962974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,16384,0.045775112178590566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,3072,0.008302222523424361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,1024,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,64,0.010379555324713389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,4096,0.045485334263907544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,1536,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,12288,0.019888889458444383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,1024,0.0038933331767717996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,12288,0.030429333448410034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,3072,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,3072,0.013462222284740873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,512,0.014678221609857349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,7168,0.03470666540993585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,10240,0.03962488969167074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,12288,0.013431111143694984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,512,0.005672889037264719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,16384,0.04272355635960897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,3072,0.007679111427730984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,65536,0.06989777750439115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,768,0.01788444485929277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,8192,0.9351563983493381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,6144,0.01824355622132619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,256,0.019906666543748643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,2560,0.010505778094132742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,128,0.009724444813198512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,1024,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,7168,0.015044444137149386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,128,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,12288,0.020819556381967332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,5120,0.011971555650234222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,7168,0.02387911081314087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,2560,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,8192,0.012488889197508493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,1536,0.008642666869693333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,128,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,2560,0.4801671240064833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,8192,0.016711999972661335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,16384,1.703266673617893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,1024,0.012774222426944308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,10240,0.028718223174413044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,1024,0.0041848888827694785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,5120,0.021244444780879553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,12288,0.02372177773051792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,2048,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,6144,0.013519110778967539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,2048,0.018952889574898612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,5120,0.014167999227841696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,4096,0.016529776983790927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,10240,0.012335110869672565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,3072,0.013450666434235044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,12288,0.015823110938072205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,256,0.002924444360865487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,3584,0.136336882909139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,6144,0.04484088884459602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,1536,0.006243555496136348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,8192,0.011100444528791638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,7168,0.27534400092230904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,7168,0.2419688966539171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,8192,0.015364444918102689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,3584,0.009168888959619734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,4096,0.06949955224990845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,6144,0.03330399923854404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,256,0.08979466888639663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,1024,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,16384,0.02508977717823452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,1024,0.014850666125615438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,2560,0.007701333198282454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,64,0.00488711107108328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,2560,0.022907555103302002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,128,0.01423733267519209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,3584,0.2660364574856228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,65536,0.08734666638904148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,10240,0.022350221872329712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,65536,0.2492977778116862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,10240,0.3761155605316162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,65536,0.03692266676161025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,3584,0.007343110938866933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,2048,0.006882666713661618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,256,0.018551111221313477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,768,0.016934222645229764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,12288,0.021514667405022517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,4096,0.008647999829716152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,8192,0.1282071140077379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,7168,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,128,0.005535111245181825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,16384,0.013376000026861826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,768,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,128,0.006232888748248418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,16384,0.02269511090384589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,1536,0.010054222411579555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,64,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,3584,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,2560,0.006304889089531369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,5120,0.01789866718981001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,7168,0.0289066665702396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,65536,0.5028657913208008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,1024,0.06289422512054443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,1024,0.0042364444169733255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,65536,0.08258755339516534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,128,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,8192,0.015623110863897534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,3584,0.01997244523631202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,4096,0.008000888758235509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,12288,0.01034755590889189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,16384,0.05357066790262858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,3072,0.04495644569396973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,16384,0.677727116478814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,3072,0.012207110722859701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,768,0.006320888797442119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,2048,0.01111288865407308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,10240,0.03093155556254917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,65536,0.06984800100326538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,512,0.005668444352017508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,256,0.009501333038012186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,65536,0.13005956013997397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,2048,0.01809155609872606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,8192,0.02423200011253357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,1024,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,3584,0.07570311095979479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,1536,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,12288,0.019352000620630052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,32,0.01035822182893753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,65536,0.07540000147289701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,8192,0.022795556320084467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,768,0.005721777677536011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,1536,0.004381333374314838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,5120,0.10570844676759507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,2560,0.0243511117166943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,8192,0.08710844649208917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,7168,0.11804889308081733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,3072,0.0901235540707906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,1024,0.024896888269318476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,5120,0.019320888651741874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,768,0.03501066565513611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,4096,0.01943733294804891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,10240,0.01536977787812551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,7168,0.06839466757244535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,3584,0.013818666338920593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,6144,0.01752355529202355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,8192,0.023278221487998962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,2560,0.00701333334048589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,32,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,768,0.003908444609906939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,1536,0.02996088729964362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,768,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,2048,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,128,0.002903999967707528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,8192,0.020250666472646926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,3584,0.026782222919993933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,768,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,3072,0.012425777812798819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,4096,0.015116444892353482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,65536,0.09272799889246623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,1024,0.009767110976907942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,12288,0.013176889055305056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,64,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,8192,0.013601777454217276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,1024,0.015586665934986539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,5120,0.006360888895061281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,6144,0.033287998702791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,3072,0.03571466604868571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,6144,0.00906399968597624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,7168,0.029824889368481104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,2560,0.05713599920272827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,10240,0.012166221936543783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,8192,0.03439555565516154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,5120,0.02211644417709774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,12288,0.013410666750537025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,64,0.003583999971548716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,5120,0.00869688888390859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,16384,0.016351110405392118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,768,0.003916444463862313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,32,0.0032808888289663526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,12288,0.007678222325113084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,7168,0.0352862212393019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,1536,0.028071999549865723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,64,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,3584,0.07246310843361749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,3072,0.010926222635640038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,12288,0.01860000027550591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,512,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,512,0.006002666635645761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,2560,0.01864177816443973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,768,0.0094035557574696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,16384,0.022942221826977197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,10240,0.13592533270517984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,2048,0.014678221609857349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,1024,0.005969777703285217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,3584,0.0081386665503184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,7168,0.01644799941115909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,128,0.014129777749379476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,3584,0.10417155424753825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,768,0.007365333537260692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,4096,0.012934221989578672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,3072,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,768,0.023374223046832617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,256,0.011880888707107969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,3584,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,3072,0.021007110675175984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,4096,0.010650667051474253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,64,0.16519822014702693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,256,0.004229333251714706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,12288,0.01748533381356133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,2048,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,12288,0.10570310884051853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,6144,0.010766222245163388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,10240,0.026007999976476032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,256,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,32,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,10240,0.015248888068728976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,3584,0.007657777931955125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,2560,0.0070675553547011475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,256,0.00563733321097162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,10240,0.0273617770936754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,5120,0.017236444685194228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,2560,0.015442666080262927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,7168,0.02718844513098399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,10240,0.03242222136921353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,32,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,128,0.0028417776856157514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,16384,0.04978222317165799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,16384,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,12288,0.020209777686331008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,2048,0.03932888971434699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,8192,0.017156443662113614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,768,0.013283555706342062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,3584,0.01789155602455139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,512,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,256,0.004184000194072723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,65536,0.08737778001361424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,5120,0.01351111133893331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,10240,0.03086577852567037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,1024,0.0038648889296584656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,3072,0.007344889144102733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,64,0.00666933341158761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,12288,0.027956446011861164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,3584,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,128,0.009866666462686326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,512,0.036325332191255354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,1536,0.024381332927280005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,3072,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,256,0.018532445033391316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,3072,0.023028444912698533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,2048,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,16384,0.02680711117055681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,512,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,128,0.0028853333658642242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,10240,0.01640266676743825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,2560,0.015447111593352424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,3072,0.019146665930747986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,768,0.016599999533759225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,768,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,512,0.010379555324713389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,1536,0.0063999998900625445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,10240,0.05348622136645847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,1024,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,3584,0.02186311119132572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,768,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,8192,0.03634488913748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,4096,0.05140622125731575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,16384,0.03872888949182298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,3072,0.007332444190979004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,64,0.005681777579916849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,3584,0.612793763478597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,12288,0.065519114335378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,1024,0.00980444418059455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,512,0.04631111025810242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,7168,0.019877332780096266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,12288,0.029764443635940552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,2560,0.024586665961477492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,128,0.018536888890796237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,2048,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,65536,0.14797777599758571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,4096,0.020008888509538438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,3584,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,10240,0.01752088963985443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,10240,0.021346666746669348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,512,0.003212444484233856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,3584,0.023747555083698694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,64,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,512,0.014880888991885714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,128,0.014463111758232117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,3584,0.009095999929640029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,12288,0.018529777725537617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,1536,0.009938666390048133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,2048,0.04782488942146301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,7168,0.023147554861174688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,6144,0.010035555395815108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,8192,0.012784000072214337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,3072,0.009217778013812171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,16384,0.02250755495495266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,16384,0.01479555500878228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,3584,0.028572443458769057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,10240,0.019612444771660697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,128,0.015560888581805758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,4096,0.00830844458606508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,8192,0.011959111524952782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,1536,0.015207999282413058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,16384,0.03135644396146139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,768,0.01661600006951226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,7168,0.009745777481132084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,7168,0.019650666250122916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,128,0.006324444380071428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,7168,0.031670222679773964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,10240,0.014340443743599786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,512,0.0461768905321757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,65536,0.20911110772026908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,32,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,6144,0.007271111011505127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,128,0.011044444309340583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,1536,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,5120,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,2560,0.006140444427728653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,64,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,8192,0.37945244047376847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,64,0.0027671110712819626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,10240,0.012800888882742988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,1024,0.008071111308203803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,128,0.0029075555503368378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,512,0.01073066641887029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,3072,0.010497777826256223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,256,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,12288,0.01900800069173177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,2048,0.008071111308203803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,2048,0.015837333268589444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,3584,0.027208000421524048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,65536,6.454329596625434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,12288,0.0745119982295566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,4096,0.013728888498412238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,32,0.004919111314747068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,32,0.028208000792397395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,2560,0.011780444118711682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,3584,0.0221760008070204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,10240,0.018312000566058688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,5120,0.024746666351954143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,12288,0.018996444013383653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,128,0.013735999663670858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,10240,0.046089778343836464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,16384,0.31282133526272243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,7168,0.022104889154434204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,6144,0.009337777892748514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,8192,0.007281777759393056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,1536,0.02682755556371477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,2048,0.00666222224632899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,7168,0.021440888444582622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,2560,0.0069466663731469055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,128,0.00832088871134652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,2560,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,3584,0.016855110724767048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,32,0.004355555607212914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,32,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,10240,0.008216000265545314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,2560,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,3584,0.021472889516088698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,512,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,768,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,512,0.004651555584536659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,4096,0.007311111523045435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,3584,0.012697777814335294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,1024,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,16384,0.020662221643659804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,256,0.015251555376582675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,3072,0.3459377818637424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,1536,0.007034666836261749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,1536,0.021144888467258875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,512,0.00564444437623024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,5120,0.14003822538587782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,8192,0.05834133095211453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,256,0.01692800058258904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,128,0.029138667715920344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,5120,0.011400889191362592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,2560,0.006670222100284364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,5120,0.10376177893744574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,4096,0.01218933363755544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,7168,0.019573332534896005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,10240,0.033703111939960055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,1024,0.006339555399285422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,64,0.005649777750174205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,10240,0.04450400008095635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,256,0.006159111029571957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,12288,0.036488887336519026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,65536,0.2421813276078966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,12288,0.018226666582955253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,2048,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,5120,0.10393510924445258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,16384,0.0611777769194709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,3584,0.01963733302222358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,3584,0.007548444800906711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,8192,0.07509777943293254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,1024,0.003924444317817688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,1024,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,5120,0.02531822191344367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,10240,0.015586665934986539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,2560,0.03414311011632284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,3584,0.10584977600309585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,6144,0.06300710969501071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,10240,0.07101333141326904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,4096,0.017239999439981248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,128,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,1536,0.004938666605287128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,1024,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,2048,0.21012355221642387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,65536,0.06665777497821383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,2560,0.007332444190979004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,4096,0.025536888175540503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,3072,0.04314488834804959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,65536,0.02380977736579047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,3072,0.011052444577217102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,5120,0.03629600008328756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,2560,0.01686044368478987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,1536,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,3072,0.016601777738995023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,256,0.007063111497296228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,256,0.0031831111345026228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,5120,0.008661333057615492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,12288,0.26390311453077525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,1024,0.007988444632954067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,2560,0.0069057775868309875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,2048,0.011044444309340583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,128,0.002999111182159848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,8192,0.014891554911931356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,2048,0.017624000708262127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,256,0.0032559999575217566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,3584,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,128,0.014544889330863953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,3584,0.013587555951542325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,65536,0.0678115553326077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,16384,0.04370489054256015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,256,0.011092444260915121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,3584,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,3072,0.008119111259778341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,512,0.007640000018808577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,16384,0.5792489051818848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,2560,0.006959111326270633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,12288,0.033816887272728816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,65536,0.059246222178141274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,32,0.004928888960017098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,2048,0.009735999835862054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,7168,0.033174223370022245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,8192,0.014675555957688225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,512,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,5120,0.018906666172875296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,3584,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,4096,0.010370666782061258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,32,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,7168,0.011436444189813403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,10240,0.016559999850061204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,4096,0.009944889280531142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,2048,0.060263997978634305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,3584,0.021963554951879714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,2048,0.01995822290579478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,3072,0.010021333065297868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,1536,0.020774222082561918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,128,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,16384,0.02479733361138238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,768,0.003530666646030214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,768,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,2048,0.1692284478081597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,8192,0.02494222256872389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,16384,0.04118488894568549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,65536,0.04765866531266106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,3072,0.007304888632562425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,2560,0.10604355732599895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,3072,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,7168,0.02344177828894721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,16384,0.10083288616604275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,1024,0.01348000019788742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,16384,0.016136000553766888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,8192,0.03840711050563388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,6144,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,4096,0.014523555835088095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,768,0.003552000141806073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,8192,0.010803555448849997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,10240,0.037476443582110934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,16384,0.039148443275027804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,3072,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,8192,0.04492888847986857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,16384,0.15052799383799234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,7168,0.026758222116364375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,3072,0.008642666869693333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,7168,0.027288888891537983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,64,0.003579555700222651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,12288,0.016531555189026725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,768,0.007833778030342525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,512,0.01350044459104538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,2048,0.0070168889231152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,128,0.006009777800904379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,128,0.010372444159454769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,256,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,10240,0.02177511155605316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,1024,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,1024,0.015104000767072042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,512,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,2560,0.010093332992659675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,4096,0.03877244393030802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,256,0.009333333207501305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,256,0.003551111039188173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,32,0.003152888889114062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,8192,0.013110222087966071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,32,0.0038737778862317405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,10240,0.007387555307812161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,512,0.01936177743805779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,16384,0.07610133621427748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,256,0.015117333994971381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,10240,0.02599999970859951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,12288,0.011035555766688453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,10240,0.03939911060863071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,12288,0.05725066529379951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,128,0.002883555574549569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,6144,0.019546666079097323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,768,0.003583111282851961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,7168,0.19384355015224883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,3584,0.012135110795497894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,2560,0.01016622202263938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,7168,0.010231999887360467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,1024,0.0170168893204795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,3584,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,2048,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,512,0.005636444522274866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,32,0.012103111379676394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,128,0.00563733321097162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,12288,0.020025778147909377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,3584,0.014489778214030795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,64,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,65536,0.25968888070848256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,3072,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,768,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,10240,0.02907199992073907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,16384,0.13342132833268908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,10240,0.014070222775141398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,2560,0.007370666497283512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,512,0.01052177780204349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,3072,0.01074577785200543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,65536,0.0865208903948466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,256,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,16384,0.15974756081899008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,768,0.0038408889538711975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,8192,0.023399111297395494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,16384,0.015826667348543804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,3584,0.04776533444722494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,128,0.013576889203654395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,768,0.007473777565691207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,1024,0.004221333397759331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,2560,0.01514933341079288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,2048,0.007502222226725684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,1536,0.005180444568395615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,3072,0.007637333538797166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,2048,0.020626667473051283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,10240,0.01699822147687276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,12288,0.011319999893506369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,3584,0.007484444313579136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,65536,0.39933154318067765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,256,0.015598222613334656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,768,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,5120,0.010056888891590966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,5120,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,128,0.005972444597217772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,1536,0.020855110552575853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,12288,0.257222228580051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,1536,0.029301332102881536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,7168,0.009672000176376766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,1024,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,12288,0.03751822312672933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,10240,0.03348977698220147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,256,0.0063573333124319715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,768,0.017785777648289997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,512,0.003169777699642711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,1024,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,1536,0.009013333254390294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,4096,0.011942221886581846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,10240,0.036525332265430026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,3072,0.0074133334888352295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,16384,0.025420443879233465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,3584,0.008001777860853408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,10240,0.06263199779722425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,32,0.005666666560702854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,6144,0.009367999931176504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,64,0.0029013332807355454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,2048,0.023992000354660883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,16384,0.07931822538375854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,7168,0.012048888537618848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,16384,0.011004444625642566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,10240,0.01997955474588606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,6144,0.018193778064515855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,10240,0.13100532690684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,3072,0.012543111211723752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,3584,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,12288,0.05523288912243313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,10240,0.04643644558058845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,2560,0.007061333292060428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,5120,0.02081155611409081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,2560,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,256,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,256,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,32,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,10240,0.12031643920474583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,512,0.005261333452330695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,12288,0.10133866469065349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,1536,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,5120,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,1536,0.024783111280865137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,256,0.011717333561844297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,3584,0.010967999696731567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,16384,0.04450133442878723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,256,0.003563555578390757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,4096,0.45810312694973415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,64,0.020044444335831534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,1024,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,3072,0.01663377715481652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,2560,0.028198222319285076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,12288,0.03195110956827799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,1536,0.008360889222886827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,512,0.011277333729796939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,2560,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,12288,0.03520088725619846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,65536,0.1329395506117079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,768,0.0069271110826068455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,256,0.015876443849669564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,1024,0.02130399975511763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,65536,0.02609155575434367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,512,0.003903111235962974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,128,0.02032088902261522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,8192,0.014683556225564746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,512,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,768,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,8192,0.021740444832377966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,6144,0.0188071115149392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,512,0.04403733213742574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,2560,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,512,0.007697777615653143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,3584,0.016607999801635742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,3072,0.02517866591612498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,16384,0.01687111126052009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,1024,0.006191111273235745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,512,0.011404444773991903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,2560,0.022203556365436975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,16384,0.37204445732964414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,3584,0.01757777730623881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,1024,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,2048,0.00795733349190818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,3072,0.1522551112704807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,32,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,128,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,512,0.005640888793600931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,2048,0.0883973307079739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,256,0.004517333375083076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,128,0.005346666607591841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,4096,0.008414222134484185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,3584,0.040941334433025785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,3584,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,768,0.00424888895617591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,1536,0.008330666356616551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,768,0.007349333001507654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,128,0.0052062223354975385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,32,0.005571555760171678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,3072,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,128,0.004261333495378494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,65536,0.09168977869881524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,8192,0.046945777204301625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,6144,0.048235555489857994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,512,0.005640000104904175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,3584,0.02058577703105079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,128,0.01513155632548862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,64,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,10240,0.02606933315594991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,16384,0.01866222255759769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,1024,0.006234666539563074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,4096,0.009777777724795872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,128,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,2048,0.030332446098327637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,12288,0.018263111511866253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,6144,0.009415999882751042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,256,0.013933334085676404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,3072,0.035087111923429705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,8192,0.02325422234005398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,5120,0.011786667009194693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,65536,0.10085599952273899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,4096,0.015826667348543804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,10240,0.008997333546479544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,4096,4.485050625271268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,128,0.009009777671760982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,2048,0.008375999828179678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,2048,0.01553066737122006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,65536,0.1483751138051351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,3584,0.031649778286616005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,256,0.0031759999692440033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,768,0.00416088890698221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,1536,0.004911111046870549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,5120,0.016520000166363187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,64,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,1536,0.06675110922919379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,12288,0.01685244507259793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,256,0.002926222152180142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,7168,0.8001982371012369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,512,0.02021422154373593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,768,0.012457778056462606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,2048,0.010416888528399997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,65536,0.0682435565524631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,1536,0.01036711119943195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,2560,0.01035822182893753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,65536,0.025366221865018208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,2048,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,5120,0.03647911217477586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,4096,0.007903111477692923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,10240,0.022106667359670002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,2048,0.007000889215204451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,6144,0.010415111151006486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,3072,0.05522488885455661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,128,0.005973333285914526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,32,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,4096,0.009018667042255402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,512,0.01886311173439026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,65536,1.3274880515204537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,256,0.027513777216275532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,2560,0.03313866588804457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,2560,0.01499288943078783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,4096,0.0151235560576121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,6144,0.11903822422027588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,7168,0.05641244517432319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,512,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,6144,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,65536,0.12007199393378364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,4096,0.01147555559873581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,3072,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,16384,0.025368889172871906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,16384,0.022315555148654517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,768,0.01884622209601932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,32,0.003198222153716617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,3072,0.01075466639465756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,5120,0.038578665918774076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,16384,0.0103502223889033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,1024,0.019973332683245342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,8192,0.0845937795109219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,6144,0.009320889082219865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,256,0.008402666283978356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,2048,0.006791111081838608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,1024,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,8192,0.03050133254792955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,5120,0.0120666664507654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,3584,0.00923733330435223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,32,0.006137777947717243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,16384,0.021495999561415777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,8192,0.04520533482233683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,8192,0.02295733326011234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,768,0.005984888722499211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,3584,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,6144,0.026206221845414903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,768,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,2048,0.007374222079912822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,7168,0.04426755507787069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,1024,0.00941244430012173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,7168,0.01826755536927117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,2560,0.015223110715548197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,256,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,10240,0.0185217774576611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,128,0.006611555400821898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,512,0.0052426668504873914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,3072,1.7374248504638672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,65536,0.035410665803485446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,16384,0.04014488723542955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,768,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,2560,0.03567377726236979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,2048,0.015791111522250705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,512,0.008467555873923832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,8192,0.026335999369621277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,128,0.06384799877802531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,8192,0.006923555499977536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,32,0.0038337777886125776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,7168,0.010206222534179688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,6144,0.020095111595259774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,10240,0.020983111527231004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,12288,0.030200888713200886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,768,0.003563555578390757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,1024,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,3584,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,3584,0.007968000239796108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,3072,0.007299555672539606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,12288,0.07559200127919515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,65536,0.07986755503548516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,768,0.016629333297411602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,128,0.010062221851613786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,32,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,12288,0.04849866694874234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,6144,0.02603111167748769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,16384,0.015665777855449252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,8192,0.011364444262451597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,8192,0.013529777526855469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,2048,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,2048,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,8192,0.01549600064754486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,3072,0.01519555515713162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,2048,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,7168,0.04532533221774631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,3584,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,12288,0.026106667187478807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,7168,0.018272888329293992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,7168,0.17635822296142578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,10240,0.15354843934377035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,3072,0.01332266628742218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,128,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,768,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,65536,0.16704977883232963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,32,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,5120,0.011786667009194693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,3072,0.009155555731720394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,2560,0.007128888534175024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,5120,0.01869333287080129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,128,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,256,0.0053377776510185665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,8192,0.3127182324727376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,4096,0.016302223006884258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,3584,0.015511110424995422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,128,0.005655111124118169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,512,0.008007999923494127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,7168,0.02053333322207133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,16384,0.037998222642474704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,768,0.006226666685607698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,2560,0.07901599672105578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,10240,0.0532862212922838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,64,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,4096,0.03594933284653558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,12288,0.011443555355072021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,10240,0.033265779415766396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,8192,0.020357333951526217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,512,0.005277333160241445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,5120,0.05532000131077237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,3072,0.06603644291559856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,256,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,3072,0.03468711177508036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,7168,0.051282667451434664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,6144,0.012424888710180918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,4096,0.01571644511487749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,2048,0.00526577772365676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,10240,0.3298319975535075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,3584,0.04575466778543261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,12288,0.5482115745544434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,6144,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,64,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,256,0.01921866668595208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,1536,0.012772444221708508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,5120,0.026752889156341553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,12288,0.01071200023094813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,16384,0.015817777978049386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,2560,0.007110222346252865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,65536,0.023294222023752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,7168,0.007888889147175683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,1536,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,10240,0.011713777979214987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,5120,0.08764888842900594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,7168,0.043896890348858304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,2048,0.018813333577579923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,2560,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,32,0.00665244460105896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,10240,0.028288000159793433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,2048,0.010040000081062317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,4096,0.016901334126790363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,16384,0.01640533407529195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,7168,0.007991111112965478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,8192,0.013286222186353473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,7168,0.04212977819972568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,2560,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,32,0.003572444534964032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,2048,0.008513777620262569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,32,0.002806222273243798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,8192,0.08809510866800944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,256,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,768,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,1024,0.005607111172543631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,2560,0.012321777641773224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,2560,0.01479466590616438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,12288,3.1852951049804688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,10240,0.02538577715555827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,64,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,12288,0.017248888810475666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,768,0.024313777685165405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,512,0.00319377767542998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,8192,0.0185253338681327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,1024,0.004517333375083076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,768,0.005974222388532426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,8192,0.013767999907334646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,3072,0.008662222160233391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,32,0.008007110820876228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,65536,0.21655110518137613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,32,0.003181333343187968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,16384,0.5244408713446723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,3584,0.00832799987660514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,7168,1.977951897515191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,12288,0.04623288909594218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,8192,0.013430222041077085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,256,0.0031795555518733132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,6144,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,65536,0.06080889039569431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,128,0.005417777846256892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,128,0.0028924445311228433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,32,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,7168,0.015237333046065437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,512,0.013045333325862885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,1536,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,128,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,10240,0.01754399968518151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,1536,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,1536,0.07049866517384847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,64,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,256,0.007995555798212687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,128,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,256,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,4096,0.009171555439631144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,65536,0.07743822203742133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,3072,0.016584888100624084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,1024,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,10240,0.025783111651738484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,1536,0.005212444398138259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,64,0.012707555459605323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,2048,0.02223377757602268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,8192,0.018556444181336295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,16384,0.015180443723996481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,128,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,128,0.002855111120475663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,768,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,6144,0.014254222313563028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,1536,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,512,0.004224888980388641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,8192,0.039206223355399236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,8192,0.011400889191362592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,5120,0.008310221963458592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,256,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,1536,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,12288,0.02492800023820665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,2048,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,64,0.003120888852410846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,32,0.002793777734041214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,1024,0.00453333349691497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,3584,0.016050666570663452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,768,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,5120,0.10978133148617214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,7168,0.036357333262761436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,512,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,3072,0.01872977779971229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,65536,0.14704977141486272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,3584,0.015840889679061044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,6144,0.011564444336626263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,5120,0.016545777519543965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,128,0.006412444429265128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,65536,0.19986044036017525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,4096,0.017238222890430026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,5120,0.010949333508809408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,8192,0.021876444419225056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,8192,0.01071911139620675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,3584,0.009546666509575313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,128,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,4096,0.01751377847459581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,1536,0.015118221441904703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,6144,0.013245333400037555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,256,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,1024,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,5120,0.012055111428101858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,8192,0.01904533306757609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,256,0.005680888891220093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,10240,0.22607644399007162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,1024,0.025685333543353613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,128,0.018964444597562153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,768,0.028703110085593328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,6144,0.01014755583471722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,6144,0.11500444677140977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,8192,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,2048,0.0859982238875495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,256,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,8192,0.013997332917319404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,65536,0.22732800907558867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,4096,0.0802124473783705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,3584,0.01072799993885888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,3072,0.03249866763750712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,256,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,512,0.003543111185232798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,2048,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,16384,0.03735288977622986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,3072,0.007346666521496243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,12288,0.10404000017378066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,2560,0.01368444412946701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,5120,0.012410666379663678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,3072,0.009904888768990835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,4096,0.04639110962549845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,512,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,4096,0.02400622268517812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,4096,0.02830044428507487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,512,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,32,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,16384,0.0218248897128635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,3072,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,10240,0.03939111034075419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,5120,0.0306915541489919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,2048,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,256,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,4096,0.08256800307167901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,3072,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,12288,0.0543137788772583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,128,0.007606222397751278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,32,0.0028106667515304354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,1024,0.004117333226733738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,12288,1.2119341956244576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,128,0.02369955513212416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,768,0.009733333355850643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,256,0.009686222506894005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,1024,0.006227555374304454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,768,0.008410666551854875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,5120,0.009466666314336989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,16384,0.03975111246109009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,5120,0.025459556115998164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,1024,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,768,0.00488800017370118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,1024,0.10412355264027913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,10240,0.011712888876597086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,65536,0.379639122221205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,512,0.030220445659425523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,16384,0.0357582237985399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,128,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,7168,0.06629688872231378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,1536,0.00758755538198683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,64,0.00490399988161193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,512,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,1024,0.009840000006887648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,1536,0.00674933319290479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,4096,0.09621866544087727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,16384,0.022450667288568284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,512,0.006326222171386083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,12288,0.021370665894614324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,256,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,3584,0.03613600134849548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,768,0.016528000434239704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,65536,0.06933866606818305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,32,0.009005332986513773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,1536,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,512,0.004927111168702443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,65536,0.4662417835659451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,512,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,4096,0.01685599982738495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,768,0.02347822156217363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,1024,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,3072,0.007297777467303806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,128,0.01716444392999013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,1024,0.004264889078007804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,16384,0.06451021962695651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,7168,0.010024000373151567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,7168,0.02052177819940779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,1536,0.032716443141301475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,3072,0.006115555349323485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,256,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,3584,0.017530666457282174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,3584,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,7168,0.01755466726091173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,4096,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,4096,0.016158221496476066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,1024,0.006239999913507038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,512,0.06588621934254964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,128,0.016166221764352586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,2048,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,1536,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,1536,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,3072,0.00741244438621733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,12288,0.013081777426931592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,4096,0.08044266700744629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,128,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,128,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,1024,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,768,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,2560,0.0808302230305142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,128,0.01384888920519087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,7168,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,4096,0.014184888866212634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,2048,0.012084444363911947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,16384,0.024244444237815008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,128,0.003539555602603488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,32,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,64,0.003842666745185852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,1024,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,768,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,1536,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,10240,0.04751733276579115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,768,0.005940444353553984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,256,0.005648889061477449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,64,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,5120,0.028199109766218398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,128,0.01479822231663598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,3584,0.008453333543406593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,8192,0.017063111066818237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,7168,0.020437333318922255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,3584,0.01811200049188402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,8192,0.16062755054897732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,2560,0.007728888756699032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,65536,0.19699288739098442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,16384,0.016914667354689706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,8192,0.010392888552612728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,8192,0.034232887956831194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,12288,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,128,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,768,0.10444088776906331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,3072,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,65536,0.06803911262088351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,16384,0.05369333426157633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,7168,0.010485333700974783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,3584,0.022667555345429316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,128,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,256,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,256,0.0203164451652103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,1536,0.007662222617202335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,3584,0.018372444642914664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,1536,0.0046880000995265115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,1024,0.011983111500740051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,5120,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,6144,0.010251555177900527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,6144,0.010699555277824402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,7168,0.017253332667880587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,5120,0.023953777220514085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,768,0.017709333035680983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,2048,0.023570666710535686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,768,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,64,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,1536,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,1024,0.04023377762900458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,1536,0.016190222567982145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,1536,0.0047591109242704184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,256,0.00351555562681622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,3072,0.007028444773621029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,7168,0.019137778215938143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,16384,0.03445777628156874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,512,0.013119999733236102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,2560,0.012970666918489667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,2560,0.01681688924630483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,65536,13.800010681152344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,4096,0.016578666037983365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,65536,1.8233680725097656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,1024,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,1536,0.014124444789356656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,512,0.008639111287064022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,256,0.006175111151403851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,1536,0.006719999843173557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,768,0.0038275557259718576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,2048,0.014779556128713818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,128,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,768,0.003903111235962974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,1536,0.006520888871616787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,32,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,65536,0.3116462230682373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,16384,0.31509333186679417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,4096,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,1536,0.007267555428875818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,32,0.00254844439526399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,256,0.0052435555391841466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,65536,0.30195289187961155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,10240,0.0874017808172438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,5120,0.028746667835447524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,1536,0.0151591118839052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,16384,0.015855110353893705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,128,0.014815110299322339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,4096,0.025272000167104933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,1024,0.027706666125191584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,5120,0.01927288870016734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,6144,0.03142222099834018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,16384,0.12972977426317003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,1024,0.03160888950030009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,8192,0.024703111913469102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,8192,0.015565334094895257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,12288,0.056185775332980685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,10240,0.01168444421556261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,4096,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,32,0.006569777925809224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,128,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,512,0.04762755499945747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,16384,0.08416622214847141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,64,0.007904888855086433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,10240,0.6034480200873481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,5120,0.01165066659450531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,1536,0.007976888782448238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,16384,0.023001778456899855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,512,0.06439733505249023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,16384,0.06817422310511272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,3072,0.06301066610548231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,7168,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,5120,0.01685333251953125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,10240,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,32,0.0038693332009845306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,256,0.0031475555151700974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,10240,0.03115644388728672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,3072,0.048897776338789195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,768,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,256,0.015481778317027621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,256,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,5120,0.02097333305411869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,10240,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,3584,0.008629333641793992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,65536,0.1398622194925944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,2048,0.011520889070298938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,8192,0.010776888993051318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,256,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,2048,0.012448000411192576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,256,0.002895111011134254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,64,0.005868444426192178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,4096,0.010427555276287926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,12288,0.01683911184469859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,4096,0.08952533536487156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,5120,0.015840000576443143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,16384,0.056594669818878174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,3584,0.0070222218831380205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,128,0.005285333428117964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,8192,0.02853155467245314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,5120,0.14448889096577963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,2048,0.017224889662530687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,16384,0.04957333207130432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,5120,0.023188443647490606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,4096,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,768,0.007340444458855524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,1024,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,7168,0.010337777435779572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,768,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,3584,0.014944000376595391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,2048,0.015309333801269531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,2560,0.013461333182122974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,1024,0.006969778074158563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,7168,0.009992889232105678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,5120,0.011867555479208628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,65536,0.09438133239746094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,5120,0.06277066469192505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,3584,0.04674222071965536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,1536,0.08920800023608738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,3584,0.03504000107447306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,128,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,1024,0.003858666867017746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,128,0.014801777071423002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,2048,0.016186666157510545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,1536,0.015568888849682279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,3072,0.03977510995335049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,512,0.01682133310370975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,16384,0.020703110429975722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,16384,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,2048,0.007864889171388414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,16384,0.016906667086813185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,2560,0.01352355546421475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,6144,0.0749795569313897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,7168,0.016229333149062265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,3584,0.00982844415638182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,7168,0.009786666267448002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,16384,0.017652443713612027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,256,0.00573244442542394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,2560,0.035395556026034884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,10240,0.011741333537631564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,3584,0.033655110332700945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,768,0.008949333594905006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,3072,0.010565333068370819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,3584,0.037885334756639265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,1536,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,256,0.01793599956565433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,2560,0.010873777998818291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,3584,0.021515554851955835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,768,0.007281777759393056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,128,0.009358222285906473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,4096,0.04383111000061035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,3072,0.017248888810475666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,10240,0.011786667009194693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,1024,0.024008888337347243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,256,0.008781332936551835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,32,0.00349866681628757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,3584,0.026130666335423786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,65536,0.07869333028793335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,1024,0.03349688980314467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,1536,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,256,0.003703111161788305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,2560,0.007039111521508958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,2560,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,12288,0.008568888737095727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,8192,0.01239733315176434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,7168,0.023775999744733173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,7168,0.013452444639470843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,256,0.015104000767072042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,2048,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,65536,3.181248982747396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,2048,0.009398221969604492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,3584,0.07113422287835015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,768,0.023216888308525085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,4096,0.11853688293033177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,3584,0.008213332957691616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,4096,0.024745777249336243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,2560,0.009567110902733272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,6144,0.04159911142455207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,2560,0.007258666886223688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,16384,0.013433777623706393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,7168,0.013841778039932251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,10240,0.011753777662913004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,3072,0.018544000056054857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,2560,0.01551822159025404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,1024,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,65536,0.07136888636483087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,65536,0.06551200151443481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,1536,0.007055111229419708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,10240,0.013544888959990608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,3584,0.007653333246707916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,128,0.006639110959238476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,2560,0.012039110892348819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,256,0.005230222311284807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,1024,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,12288,0.007662222617202335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,5120,0.01631644368171692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,32,0.0028568889117903183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,5120,0.01387111097574234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,3072,0.06750933329264323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,10240,0.016160888804329764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,16384,0.036580443382263184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,5120,0.013806222213639153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,2560,0.007227555745177799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,6144,0.028592887851927016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,3584,0.05095022254519992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,4096,0.008566222257084316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,3584,0.017857778403494094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,6144,0.020284444093704224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,65536,0.32283910115559894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,6144,0.008391111261314815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,16384,0.057010663880242236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,512,0.018585777944988675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,2560,0.010265777508417765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,128,0.0034959999223550162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,256,0.012455110748608908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,16384,0.032885332902272545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,4096,0.056348443031311035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,768,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,6144,0.09547111060884263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,64,0.003539555602603488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,256,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,5120,0.01888177792231242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,1024,0.014149333039919535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,3072,0.007716444631417592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,10240,0.07989244328604804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,768,0.005869333528810077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,32,0.0028595555987623003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,2048,0.01963911122745938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,1536,0.006877333339717653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,64,0.0026604444202449587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,10240,0.02494311167134179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,3072,0.009321778184837764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,6144,0.01182933317290412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,4096,0.025405334101782903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,128,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,1536,0.005973333285914526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,256,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,512,0.01927199959754944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,16384,0.01811022228664822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,2560,0.014752000570297241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,3072,0.010089777410030365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,2048,0.017822222577200997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,8192,0.01588711142539978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,3584,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,3584,0.18231733640034994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,10240,0.028522666957643297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,4096,0.008336889247099558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,32,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,65536,0.10168800089094375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,16384,0.1719342205259535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,1024,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,12288,0.01757600075668759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,128,0.0063528890411059065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,2560,0.015096000499195524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,2560,0.01683911184469859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,256,0.01686044368478987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,2048,0.01605333387851715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,768,0.015312888556056552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,512,0.017159110969967313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,3072,0.0169804443915685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,32,0.00700444479783376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,2560,0.010043555663691627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,8192,0.008012444608741337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,10240,0.03163111209869385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,6144,0.030607998371124268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,512,0.005650666852792104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,256,0.0053022222386466126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,2048,0.02288355595535702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,5120,0.031120889716678198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,16384,0.025171554750866357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,16384,0.04283377859327528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,32,0.00629333323902554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,64,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,1536,0.01995644470055898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,7168,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,32,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,3072,0.02260622216595544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,3072,0.01075644459989336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,8192,0.05548622210820516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,5120,0.02926666537920634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,2560,0.0070168889231152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,3584,0.010658666491508484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,4096,0.019353777170181274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,768,0.007616889145639207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,2560,0.013807111316257052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,7168,0.05745777818891737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,1536,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,3584,0.008046222229798635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,1024,0.0038968887594011095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,6144,0.022674666510687932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,4096,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,65536,0.014787554740905762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,128,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,12288,0.03273333443535699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,256,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,256,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,12288,0.08931822246975368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,128,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,12288,0.15473243925306532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,8192,0.01646844380431705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,6144,0.009687110781669617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,8192,0.02197244432237413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,6144,0.019401777121755812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,1536,0.023762666516833838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,10240,0.014010666145218743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,64,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,4096,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,768,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,65536,0.2896711031595866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,4096,0.029818667305840388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,4096,0.012087110843923358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,1536,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,512,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,64,0.006305777778228124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,1024,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,7168,0.009648000200589498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,8192,0.05144800080193413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,6144,0.06250488758087158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,3072,0.17286400000254312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,8192,0.016536888149049547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,7168,0.010033778018421596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,12288,0.013060444758998023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,7168,0.019719110594855416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,1536,0.015849777393870883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,10240,0.02546399997340308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,3584,0.0701226658291287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,2560,0.026408889227443274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,16384,0.03511200017399258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,2560,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,3072,0.008486222061845992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,128,0.015190222197108798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,128,0.014440889159838358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,128,0.005669333454635408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,2048,0.006527110934257507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,1024,0.004372444417741564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,2560,0.015986666083335876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,4096,0.00869599978129069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,32,0.0034506666577524612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,7168,0.551928891075982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,64,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,8192,0.054083555936813354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,32,0.004318222403526306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,128,0.007744000189834171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,2560,0.03682044479582045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,2048,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,3072,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,3072,0.020441777176327176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,6144,0.11132621765136719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,65536,0.048287110196219556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,1024,0.025400888588693406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,10240,0.02195466723706987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,64,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,2048,0.008299555215570662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,1536,0.005208000126812193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,1536,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,4096,0.013801777528391944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,16384,0.02481599979930454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,256,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,2560,0.014868444866604276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,256,0.0034933334423436057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,7168,0.013460444079505073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,3584,0.010732444624106089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,16384,0.14231022198994955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,512,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,3584,0.008708444734414419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,3072,0.00999022192425198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,6144,0.06339466571807861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,256,0.07069866524802314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,128,0.06012622515360514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,8192,0.23363733291625977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,1536,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,3072,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,65536,0.17761688762240938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,12288,0.014840000205569796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,6144,0.011289777855078379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,2560,0.009186666872766282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,5120,0.009418666362762451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,128,0.0028968888024489083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,2048,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,128,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,32,0.014827556080288358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,10240,0.10053600205315484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,2048,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,1536,0.014519999424616495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,2048,0.037068443165885076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,4096,0.008392888638708327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,512,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,256,0.014836443795098199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,16384,0.0324888891643948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,16384,0.013201778133710226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,65536,0.4535973336961534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,8192,0.01478488908873664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,256,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,768,0.017198221551047433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,12288,0.040391110711627536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,12288,0.017880888448821176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,3584,0.011083555718262991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,12288,0.0199306673473782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,256,0.06637510988447401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,512,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,8192,0.023619555764728125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,8192,0.010416000253624385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,65536,0.07368889119890001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,32,0.009840889109505547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,1024,0.007249777515729268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,65536,0.06694489055209689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,6144,0.009339555270142024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,5120,0.06565688716040717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,16384,0.043455110655890576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,2560,0.07167377736833361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,7168,0.012865777644846173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,1536,0.01348533315791024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,10240,0.11857244703504775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,10240,0.0348693331082662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,256,0.008194666769769456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,1536,0.019336889187494915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,7168,0.023023999399609033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,512,0.003169777699642711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,128,0.018127111925019156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,3072,0.017310221989949543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,65536,0.8129057884216309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,3072,0.014102222190962898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,6144,0.008971555365456475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,7168,0.04169244567553202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,128,0.0035084444615576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,128,0.005972444597217772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,2048,0.007674666742483775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,768,0.003519999898142285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,7168,0.022097777989175584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,16384,0.04853244291411506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,8192,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,7168,0.014555555250909595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,2560,0.0147733340660731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,12288,0.43081778950161403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,512,0.011424889167149862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,4096,0.00758755538198683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,3072,0.07379377550548978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,10240,0.16184533966912165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,16384,0.044567998912599355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,7168,0.018512000640233357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,2560,0.016022221909628976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,6144,0.009119999905427298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,1536,0.008969777988062965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,32,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,512,0.01314133322901196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,4096,0.017492444978819955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,8192,0.014650666051440768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,1024,0.01235466698805491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,1024,0.016900445024172466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,7168,0.09603910975986057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,16384,0.01924888955222236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,1024,0.06326488653818767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,5120,0.009015999734401703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,5120,0.02595377796226078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,8192,0.010624888870451184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,7168,0.011754666765530905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,32,0.005982222242487802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,7168,0.4935590955946181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,4096,0.022656889425383672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,65536,0.8112311363220215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,1536,0.008309333688682979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,2048,0.00833422193924586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,256,0.005607111172543631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,16384,0.04901422063509623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,7168,0.0103502223889033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,4096,0.007608888877762689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,768,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,512,0.0558240016301473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,2048,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,256,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,1024,0.02290311124589708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,128,0.02788089050187005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,4096,0.028656889994939167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,5120,0.011701333026091257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,4096,0.04068444503678216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,1024,0.02479644450876448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,65536,0.10049600071377224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,10240,0.012095111111799875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,128,0.011077333655622272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,3584,0.014156444205178155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,7168,0.01348622226052814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,6144,0.016991110311614144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,256,0.0031555555760860443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,8192,0.0863484475347731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,3584,0.007687999970383114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,16384,0.032215111785464816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,768,0.0400257772869534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,2048,0.04495111107826233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,1536,0.023245332969559565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,32,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,16384,0.05263288815816244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,128,0.011477332976129321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,1024,0.006115555349323485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,512,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,6144,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,4096,0.011371555427710215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,32,0.01037599974208408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,512,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,3584,0.008662222160233391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,6144,0.01108088923825158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,256,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,32,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,3584,0.009877333210574256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,8192,0.016015999846988253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,10240,0.011887110769748688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,4096,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,1536,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,768,0.005631999837027655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,1536,0.008375110725561777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,768,0.006633777585294511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,5120,0.020270221763186984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,768,0.004924444274769889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,12288,0.013407111167907715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,7168,0.02937600016593933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,6144,0.01872977779971229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,32,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,1536,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,3584,0.01626044511795044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,64,0.011674666570292579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,4096,0.011958222422334882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,12288,0.06112088759740194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,7168,0.03062844607565138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,16384,0.04431466592682732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,2560,0.019711999429596793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,3072,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,3584,0.014492443866199918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,12288,0.012359110845459832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,10240,0.022061333060264587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,3584,0.015819556183285184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,16384,0.026514665948020086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,16384,0.15569421980116102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,1024,0.013451555536852943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,6144,0.011240888800885944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,3584,0.030023998684353296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,128,0.007074666519959767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,2560,0.011094222466150919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,32,0.006052444378534953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,2560,0.006713777780532837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,6144,0.1403004460864597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,2048,0.014105778601434497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,2560,0.006699555450015598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,16384,0.11547377374437119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,1024,0.00599644457300504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,3072,0.012129777835475074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,8192,0.02715822226471371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,2560,0.019433778193261888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,1536,0.005241777747869492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,768,0.006305777778228124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,5120,0.01257066677014033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,128,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,65536,0.12591289149390325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,16384,0.04585955540339152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,1536,0.015258666541841296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,32,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,768,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,3072,0.017198221551047433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,16384,0.03227644496493869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,6144,0.02165866725974613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,8192,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,256,0.006719999843173557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,512,0.006705777926577463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,6144,0.024767110745112102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,7168,0.012993777791659037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,3072,0.0073315559162033936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,16384,0.04632444514168633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,1536,0.004512888689835866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,3072,0.1254862282011244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,1024,0.009460444251696268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,5120,0.016912889149453905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,3584,0.03256888853179084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,16384,0.0942995548248291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,7168,0.09017244312498306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,5120,0.01646933290693495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,2048,0.006447111152940326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,8192,0.013320888910028668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,768,0.003560888684458203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,16384,0.06931377781762017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,6144,0.02609155575434367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,2560,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,1536,0.004927111168702443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,10240,0.0388302206993103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,768,0.034147553973727755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,65536,0.0652151107788086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,4096,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,10240,0.024694222542974684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,512,0.0039022221333450745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,12288,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,6144,0.020794666475719877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,4096,0.04115555683771769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,3584,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,256,0.016899555921554565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,65536,0.2477502293056912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,65536,0.024736000431908503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,16384,0.02256444427702162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,768,0.021353777911927965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,3584,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,3584,0.010426667001512317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,128,0.0029075555503368378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,64,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,3072,0.008994667066468133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,5120,0.008416888614495596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,2048,0.02944355540805393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,512,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,512,0.004472889006137848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,10240,0.08285955588022868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,3584,0.010648888846238455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,7168,0.26116622818840873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,16384,0.05556533071729872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,1536,0.004577777865860197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,128,0.0052897776994440295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,1536,0.017720000611411203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,1024,0.0284222231970893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,4096,0.015196444259749519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,8192,0.020561777883105807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,6144,0.031135112047195435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,7168,0.01959377858373854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,2560,0.007232888705200619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,768,0.023225777679019507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,6144,0.009992889232105678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,3072,0.011755555868148804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,10240,0.030577778816223145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,2560,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,64,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,64,0.0028453332682450614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,128,0.01515111161602868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,768,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,512,0.005640000104904175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,3584,0.01386222243309021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,64,0.00300799993177255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,8192,0.361041784286499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,16384,0.022636445032225713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,6144,0.021612443857722815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,10240,0.014936000108718872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,768,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,8192,0.033765332566367254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,3584,0.00924622184700436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,8192,0.013941332697868347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,4096,0.01723377737734053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,768,0.009867555565304225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,6144,0.03334666623009576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,3584,0.015236443943447538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,768,0.0058124446206622655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,512,0.09316444396972656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,8192,0.05311377843221029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,2560,0.011767999993430244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,3072,0.013757333159446716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,4096,0.0052977779673205475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,2048,0.0053119998839166425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,1536,0.01924000018172794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,12288,0.0355022218492296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,4096,0.012837332983811697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,256,0.015481778317027621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,12288,0.15944000085194907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,3584,0.008902221918106079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,64,0.006196444647179709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,5120,0.009200889203283522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,2048,0.008629333641793992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,256,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,1536,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,10240,0.02830044428507487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,16384,0.06004710992177328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,128,0.002867555452717675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,2048,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,3072,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,1024,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,3584,0.012429333395428128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,256,0.01352266636159685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,4096,0.02107288936773936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,64,0.006069333189063602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,3584,0.026196445027987163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,5120,0.009004444711738164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,1536,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,32,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,8192,2.17191653781467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,128,0.027116444375779893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,1024,0.009047110875447592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,8192,0.015465777781274585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,2560,0.04930222365591261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,1536,0.007616889145639207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,3072,0.019719999697473314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,10240,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,768,0.005666666560702854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,2048,0.008349333372380998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,256,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,256,0.003526222374704149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,64,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,4096,0.014836443795098199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,16384,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,5120,0.01739022301303016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,7168,0.02774488925933838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,10240,1.060384856330024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,1536,0.009030222064918941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,2048,0.009678222239017487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,128,0.004579555657174853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,3072,0.019911110401153564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,7168,0.06722755564583673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,6144,0.028134223487642076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,10240,0.014841778410805596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,768,0.018910222583346896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,256,0.03461422191725837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,16384,0.8445902400546603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,5120,0.026156443688604567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,5120,0.02383377816942003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,7168,0.04120533333884345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,2560,0.02438933319515652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,512,0.025067556235525344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,3072,0.02049066623051961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,1536,0.013099555340078143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,128,0.023048889305856492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,2048,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,12288,0.013425778183672162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,1536,0.006606222026877933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,16384,0.0872195561726888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,512,0.013922666509946188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,256,0.056129780080583364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,10240,0.021376889612939622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,6144,0.008970666262838576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,5120,0.02778488728735182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,12288,0.013746666411558786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,2560,0.022795556320084467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,1536,0.052800890472200185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,768,0.016231111354298063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,6144,0.01368977791733212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,1536,0.025294222765498694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,128,0.0027475555737813315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,2048,0.007321777443091075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,6144,0.009319110876984065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,7168,0.026628444592158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,7168,0.038929777012930974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,512,0.02169955604606205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,3072,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,10240,0.012096000214417776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,2048,0.02346933384736379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,6144,0.011541333463456897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,16384,0.035776889986462064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,512,0.019952888290087383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,256,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,2560,0.05147288905249702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,4096,0.011047110789351992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,1024,0.0042071110672420925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,8192,0.02720355490843455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,4096,0.008531555533409119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,2560,0.008299555215570662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,512,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,3584,0.007440889047251807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,1536,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,12288,0.02440533373090956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,2048,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,768,0.008351999852392409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,3584,0.08803021907806396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,128,0.012086222569147745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,512,0.0127688886390792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,5120,0.040976001156700984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,256,0.012095111111799875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,1024,0.005200888961553574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,512,0.015982222225930955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,16384,0.012091555529170565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,3584,0.044714666075176664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,65536,0.13780532942877874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,64,0.006528888725572162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,3584,0.007807110746701558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,3584,0.02446577780776554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,64,0.0027831111931138565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,768,0.0058755555914507965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,3584,0.0074142225914531285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,1024,0.017859554953045316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,1536,0.014051554931534661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,16384,0.025411556164423626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,1024,0.03372888763745626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,8192,0.036370668146345354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,128,0.007691555553012424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,256,0.012426666915416718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,1024,0.007272889216740926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,128,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,10240,0.01481599940194024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,2560,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,10240,0.23728977309332955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,6144,0.01591022147072686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,8192,0.07495111227035522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,128,0.014446222119861178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,8192,0.038034667571385704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,2560,0.023544889357354905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,2560,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,8192,0.011665778027640449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,512,0.005687110953860813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,3072,0.01585244470172458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,3072,0.025768000218603346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,2560,0.016904888881577384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,4096,0.04356888930002848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,6144,0.07777866390016344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,2048,0.01682044400109185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,768,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,4096,0.008759111166000366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,4096,0.0079644446571668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,2560,0.00833422193924586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,10240,0.012479999827014076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,768,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,768,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,8192,0.03561866614553664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,3584,0.02492622203297085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,1536,0.013432888521088494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,512,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,5120,0.028428445259730022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,64,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,1024,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,512,0.004943999979231092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,65536,0.05808533562554253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,6144,0.017163554827372234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,8192,0.031888888941870794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,256,0.01920000049802992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,128,0.0029066666546795103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,1024,0.003891555385457145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,7168,0.14017332924736872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,256,0.015005333556069268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,16384,0.021070222059885662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,4096,0.009232000344329411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,128,0.20143377780914307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,4096,0.01867555578549703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,4096,0.010823999842007955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,512,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,128,0.014502222339312235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,768,0.02722666660944621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,4096,0.008647110727098253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,2048,0.02436711059676276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,512,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,3584,0.006679111056857639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,16384,0.01719911065366533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,1024,0.005056888692908817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,2560,0.020879111356205408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,4096,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,3072,0.012694222231705984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,7168,0.013002666334311167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,4096,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,5120,0.07833866940604316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,64,0.005521777603361342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,65536,0.07819733354780409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,2048,0.015370666980743408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,7168,0.008776889079146916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,256,0.009199110998047723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,5120,0.00811111099190182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,12288,0.032594667540656194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,16384,0.052559110853407115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,2560,0.014584889014561972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,3584,0.012157333393891653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,2560,0.010159999959998662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,768,0.007345777418878343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,65536,0.0684408876630995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,2048,0.09104266431596543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,2048,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,6144,0.024028443627887305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,3584,0.007780444290902879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,768,0.0038533334930737815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,4096,0.01479288935661316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,10240,0.058641778098212353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,3072,0.007631111476156447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,128,0.015797333584891427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,3072,0.019533332851197984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,16384,0.06915911038716634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,10240,0.27323465877109104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,768,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,7168,0.012464000119103326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,768,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,3072,0.010040000081062317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,256,0.028071999549865723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,1536,0.017906667457686532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,10240,0.02272622287273407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,1536,0.023558222585254248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,12288,0.016163556112183463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,32,0.02232444451914893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,4096,0.010426667001512317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,512,0.0038373333712418876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,3072,0.01732711162832048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,8192,0.04541155695915222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,10240,0.037069333924187556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,1024,0.009204444785912832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,512,0.005722666780153911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,1536,0.004522666749027041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,65536,9.280392116970486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,768,0.0069271110826068455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,3072,0.007243555453088548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,16384,0.4050390985276964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,12288,0.2648986710442437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,256,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,1536,0.010991999672518836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,6144,0.024496000674035814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,6144,0.012727110750145383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,3072,0.018790221876568265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,3072,0.008651555412345463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,768,0.007419555551475949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,256,0.026769777139027912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,1536,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,16384,0.04128977656364441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,10240,0.05660177601708306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,64,0.010132444401582083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,3584,0.009385777844323052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,12288,0.013649777405791812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,65536,0.04558044340875414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,1536,0.04465511110093859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,5120,0.00868800034125646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,256,0.021057777934604224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,1536,0.006562666760550604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,1536,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,16384,0.04192622171507942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,512,0.004599110947714912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,12288,0.013408000270525614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,2560,0.007375111182530721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,2560,0.006907555378145642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,2560,0.02833599845568339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,2048,0.143230226304796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,32,0.006405333264006509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,5120,0.2712026702033149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,128,0.003157333367400699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,10240,0.03443822264671326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,7168,0.020581333173645865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,8192,0.011879111329714457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,1024,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,1536,0.016472889317406546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,10240,0.012423111332787408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,6144,0.10083822409311931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,1024,0.08570310804578994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,3072,0.010741333166758219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,5120,0.010610666539933948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,7168,0.027801778581407335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,128,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,512,0.004201777693298128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,65536,0.05785422192679512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,128,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,2048,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,1536,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,8192,0.010681777364677854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,5120,0.018175111876593698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,8192,0.01870666609870063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,16384,0.044221334987216525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,2048,0.0070088886552386815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,16384,0.009664889011118147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,768,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,7168,0.017623111605644226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,1536,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,7168,0.12631555398305258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,256,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,256,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,32,0.0028924445311228433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,2560,0.011397333608733283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,128,0.005636444522274866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,512,0.0038853333228164246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,64,0.0032382222513357797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,65536,0.0692231125301785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,3072,0.015463110473420886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,5120,0.08068711227840848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,64,0.0032035555276605818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,6144,0.030601776308483545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,512,0.01628266606065962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,1536,0.009012444151772393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,128,0.056758224964141846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,128,0.006909333169460297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,7168,0.1037751105096605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,6144,0.009674666656388177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,6144,0.018350222044520907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,12288,0.01959377858373854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,1024,0.0042275554604000514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,32,0.0028568889117903183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,768,0.018713777263959248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,2048,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,64,0.0035786665976047516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,4096,0.011369778050316704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,2560,0.022346667117542688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,128,0.015136889285511441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,2048,0.009712888962692684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,128,0.011224889092975192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,128,0.013196444345845116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,256,0.005440000030729506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,8192,0.010491555763615502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,1536,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,256,0.004226666771703296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,128,0.010394666757848527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,1024,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,10240,0.04403822289572822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,3072,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,512,0.005879999862776862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,7168,0.011757333245542316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,4096,0.008056888977686564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,3584,0.012155555188655853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,512,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,3584,0.011709333293967776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,4096,0.012324444121784635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,5120,0.011380444798204633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,2560,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,65536,0.13679644796583387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,4096,0.016869333055284288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,12288,0.01648000048266517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,2560,0.006917333437336816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,2048,0.02603111167748769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,512,0.003859555555714501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,256,0.024106666445732117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,2560,0.032264000839657254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,12288,0.036200000180138484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,5120,0.020445333586798776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,12288,0.05446488989724053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,1024,0.018921777606010437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,2048,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,12288,0.04831022355291578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,3584,0.012868444124857584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,2048,0.014446222119861178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,2560,0.016359999775886536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,512,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,12288,0.016136000553766888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,65536,0.08159022198783027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,1024,0.008025777836640676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,5120,0.014004444082578024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,16384,0.028759999407662287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,1024,0.02184444500340356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,512,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,16384,8.93115488688151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,16384,0.03524533245298598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,5120,0.023780445257822674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,7168,0.10327466991212632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,7168,0.028395556741290625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,2560,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,3072,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,2048,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,8192,4.105951097276476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,4096,0.05444000164667765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,7168,0.08389777607387966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,8192,0.010410666465759277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,2560,0.007538666327794393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,10240,0.009378666679064432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,512,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,1536,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,10240,0.02730577852990892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,12288,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,256,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,6144,0.040264000495274864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,5120,0.00904799997806549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,3072,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,3072,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,1024,0.013736888766288757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,3584,0.023774221539497375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,12288,0.021527111530303955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,12288,0.06159733401404487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,12288,0.04371644390953911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,1024,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,128,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,512,0.005680000202523337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,16384,0.03127022253142463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,65536,0.015126221709781222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,3072,0.026138666603300307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,1536,0.004566222015354368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,2048,0.05522666705979241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,12288,0.35671289761861164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,256,0.024673778149816725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,3072,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,4096,0.008012444608741337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,65536,0.17098400327894422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,256,0.017502221796247695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,7168,0.014179555906189812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,1024,0.013075555364290873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,1024,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,3072,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,512,0.004176889028814104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,4096,0.027899556689792212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,128,0.0064222220745351576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,128,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,8192,0.026151999831199646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,6144,0.018290667070282828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,3584,0.007869333028793335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,3072,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,12288,0.016499555773205228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,65536,0.19009155697292754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,65536,0.20606666141086152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,256,0.01019022199842665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,65536,0.060905781057145864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,2048,0.049903111325369946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,3072,0.010056888891590966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,8192,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,32,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,768,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,128,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,3584,0.021839999490314062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,8192,0.015056888262430826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,10240,0.049245334333843656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,64,0.005964444329341252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,512,0.005333333379692501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,1536,0.019540444016456604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,3584,0.030878222650951807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,128,0.01534844438234965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,10240,0.021664000219768945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,3072,0.007048889166778988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,1024,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,2048,0.09954666429095799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,7168,0.07834666967391968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,128,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,512,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,512,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,3072,0.00702044450574451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,65536,0.14585955937703451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,8192,0.024320888850424025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,5120,0.06669333246019152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,512,0.0053191110491752625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,256,0.006508444332414203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,128,0.015568888849682279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,1536,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,3072,0.02104266650146908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,5120,0.009015999734401703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,128,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,12288,0.04450044367048475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,10240,0.02962577674123976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,3584,0.007312888900438945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,8192,0.030561778280470107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,6144,0.010281778044170803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,256,0.0052586665583981415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,3584,0.01036266651418474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,256,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,1536,0.016530666086408827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,128,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,1536,0.018886221779717337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,65536,0.09865422381295098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,2560,0.009005332986513773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,8192,0.023557333482636347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,1024,0.005717333406209946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,7168,0.03610488772392273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,3072,0.025507556067572698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,64,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,1024,0.014138667119873894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,5120,0.008938666847017076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,5120,0.11808266904619004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,1536,0.013460444079505073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,768,0.0042408886882993905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,2048,0.013816888961527081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,32,0.0053182223604785065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,8192,0.014488889111412896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,128,0.014446222119861178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,1024,0.01788444485929277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,8192,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,2048,0.015489778584904142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,3072,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,65536,0.07381866375605266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,2560,0.008063111040327284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,12288,0.012103999654452005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,7168,0.03347999850908915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,12288,1.0601200527615018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,16384,0.03421511252721151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,2560,0.007251555720965068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,768,0.006731555693679386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,4096,0.007657777931955125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,1024,0.01791466606987847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,32,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,3072,0.027448000179396734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,6144,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,5120,0.019882667395803664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,16384,0.04469688733418783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,512,0.006520000182920032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,128,0.0028231110837724474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,3584,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,2048,0.007259555160999298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,12288,0.039056890540652804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,4096,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,6144,0.008956444760163626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,32,0.003220444545149803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,768,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,1024,0.006190222170617845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,2048,0.006619555668698416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,2560,0.007635555333561367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,6144,0.18459288279215494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,2560,0.0281057788266076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,256,0.010031999813185798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,4096,0.03696799940533108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,10240,0.695531580183241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,512,0.011011555790901184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,3072,0.0165964447789722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,768,0.017804443836212158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,3072,0.008202667037645975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,6144,0.1158755620320638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,3584,0.010296888649463654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,8192,0.026437333888477747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,10240,0.03336800138155619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,4096,0.01703555550840166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,4096,0.08054222000969781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,1024,0.034113778008355036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,4096,0.01555288831392924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,5120,0.01090222183201048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,64,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,3072,0.02808622188038296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,256,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,512,0.022336888644430373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,128,0.0052737775776121355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,32,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,3072,0.015136000182893542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,16384,0.023891554938422307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,6144,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,768,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,2560,0.0075902218619982404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,6144,0.029149333635965984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,3072,0.017868444323539734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,10240,0.03042844600147671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,6144,0.017833777599864535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,10240,0.05465333329306709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,8192,0.01034222212102678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,1536,0.009142222503821055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,768,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,1536,0.01145333300034205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,65536,0.13534666432274714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,65536,0.3882577684190538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,768,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,65536,0.12342311276329888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,512,0.016560888952679105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,5120,0.22472532590230307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,1024,0.027090667022599116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,512,0.005669333454635408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,768,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,512,0.010036444498433007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,4096,0.009091555244392818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,5120,0.008719999757077958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,2560,0.008022222254011366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,128,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,12288,0.02847911251915826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,768,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,4096,0.10645777649349636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,512,0.00425511101881663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,512,0.018569777409235638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,5120,0.01721155477894677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,128,0.010122666756312052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,2048,0.008270222279760573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,1024,0.007339555356237624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,2048,0.023989333046807185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,8192,0.014519110321998596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,32,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,3072,0.015445333388116626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,256,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,12288,0.014871999621391296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,2560,0.006217777729034424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,3072,0.016581333345837064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,65536,0.015114666687117683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,4096,0.018970666660202872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,2560,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,3584,0.007320000065697565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,128,0.0058702222175068325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,16384,0.02725155486000909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,3072,0.013857777747843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,10240,0.007689778175618913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,128,0.005230222311284807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,3072,0.00885511106914944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,16384,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,5120,0.10522933138741387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,10240,0.032241778241263494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,7168,0.019917334119478863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,6144,0.01870133313867781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,3584,0.010332444475756751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,64,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,10240,0.046967999802695386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,3584,0.007340444458855524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,7168,0.015089778436554803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,16384,0.5294035805596246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,2048,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,1024,0.007608888877762689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,768,0.01074488874938753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,8192,0.028027554353078205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,512,0.00490488898422983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,768,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,8192,0.021477333373493616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,768,0.0038577777643998465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,1024,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,6144,0.05651288562350803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,256,0.0031679999083280563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,3584,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,65536,0.06203022268083361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,64,0.0028417776856157514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,2560,0.007685333490371704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,1024,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,32,0.002801777794957161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,12288,0.016712889075279236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,3072,0.0070773329999711775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,12288,0.021307556165589228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,4096,0.009687110781669617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,12288,0.045891556474897593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,5120,0.014913777510325113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,4096,0.014132445057233175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,4096,0.022322666313913133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,3072,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,512,0.006726222319735422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,16384,0.039552887280782066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,10240,0.02130399975511763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,512,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,5120,0.010406222608354356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,32,0.002576888849337896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,8192,0.014742222097184924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,1536,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,128,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,512,0.0032115555885765287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,256,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,3584,0.010210666391584609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,7168,0.019845333364274766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,65536,0.13553333282470703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,64,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,512,0.003314666656984223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,3584,0.009922666682137383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,3072,0.010857777463065254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,5120,0.019959999455346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,6144,0.01034577770365609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,65536,0.2851288848453098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,1536,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,16384,0.1938880019717746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,1024,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,256,0.003887999802827835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,1024,0.15662044949001735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,1024,0.01713244451416863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,4096,0.008273777862389883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,6144,0.008365333080291748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,7168,0.11932533317142063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,128,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,7168,0.011408000356621213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,2048,0.023932443724738225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,65536,0.09438666370179917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,10240,0.033688889609442815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,6144,0.009745777481132084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,65536,0.008985777695973715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,2048,0.40073156356811523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,5120,0.016869333055284288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,16384,0.06841244300206502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,12288,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,768,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,512,0.003488000068399641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,6144,0.12819022602505153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,65536,0.06710311439302233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,1024,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,1024,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,512,0.0034799998005231223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,3072,0.0824142230881585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,6144,0.03663466705216302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,1024,0.006909333169460297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,1024,0.01423911088042789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,5120,0.0313991109530131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,8192,0.014204444156752693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,4096,0.008010666403505538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,2048,0.022925333844290838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,768,0.0151555554734336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,4096,0.007606222397751278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,3584,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,128,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,768,0.03172355559137132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,128,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,7168,0.012520888613329994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,4096,0.12609867254892984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,7168,0.021558221843507554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,128,0.005232000102599462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,1024,0.01623288955953386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,4096,0.014906666345066495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,65536,0.0710995528433058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,768,0.007965332931942409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,2560,0.022016889519161646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,3584,0.006973333656787872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,6144,0.27890221277872723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,3072,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,2560,0.006609777609507243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,3072,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,6144,0.0543066660563151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,65536,1.6643804974026148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,4096,0.0492835541566213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,2048,0.05966222286224365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,7168,0.05724177757898966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,1024,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,512,0.016196444630622864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,32,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,8192,0.20541244082980684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,5120,0.03868088788456387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,1024,0.012664889295895895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,5120,0.01959022217326694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,10240,0.031311998764673867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,10240,0.02291733357641432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,16384,0.04326755470699734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,10240,0.440062231487698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,1024,0.004868444469239977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,12288,0.06466222471661039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,2560,0.014866666661368476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,2048,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,512,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,16384,0.04236266679233975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,1024,0.011976888610257043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,6144,0.012100444071822695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,2048,0.013728888498412238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,16384,0.08783288796742757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,10240,0.011050666371981302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,2560,0.034344888395733304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,1024,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,10240,0.06854933500289917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,1024,0.017483555608325534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,4096,0.010661332971519895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,128,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,2560,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,4096,0.026390223039521113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,16384,0.1745040019353231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,10240,0.02956355611483256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,3584,0.010781333678298526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,2048,0.006566222343179915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,6144,0.0096124443742964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,512,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,512,0.009015111459626092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,5120,0.013787555197874704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,6144,0.08309600088331434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,4096,0.02553333342075348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,12288,0.04945955673853556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,64,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,256,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,8192,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,32,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,64,0.08079822195900811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,4096,0.009319999979601966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,10240,0.10352533393436009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,1024,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,16384,0.021725333399242822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,32,0.003165333428316646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,6144,0.15557600392235651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,65536,0.12547289000617132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,16384,0.02400355537732442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,10240,0.012890666723251343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,3584,0.219777782758077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,32,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,5120,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,1024,0.01035733355416192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,2560,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,10240,0.10029866960313584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,7168,0.008008889026112026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,4096,0.008649778034951951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,768,0.014163555370436775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,4096,0.01590844492117564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,256,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,6144,0.009678222239017487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,4096,0.06357866525650024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,8192,0.01923111081123352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,12288,0.016689777374267578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,4096,0.01886399918132358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,1536,0.004582222137186262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,10240,0.012420444852775998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,65536,0.1263128916422526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,128,0.01482488877243466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,3584,0.16844356060028076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,7168,0.014276444911956787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,7168,0.03858133488231235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,3584,0.007613333563009898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,1536,0.19028355015648735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,512,0.017599110802014668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,16384,0.14139733049604628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,768,0.00702311098575592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,2048,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,16384,0.030638221237394545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,12288,0.0425911115275489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,65536,0.061053335666656494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,2560,0.025802666942278545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,8192,0.016190222567982145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,16384,0.2951635519663493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,12288,0.016176000237464905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,32,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,1536,0.08873066637251113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,65536,0.13785689406924778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,3584,0.08832355340321858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,2048,0.060080885887145996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,5120,0.026592888765864905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,3072,0.016146666473812528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,32,0.002493333278430833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,768,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,7168,4.346721649169922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,256,0.005256888767083486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,256,0.022310222188631695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,12288,0.061887111928727895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,4096,0.020741333564122517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,7168,0.023019555542204116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,5120,0.03702400128046671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,2048,0.038789331912994385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,1024,0.019271110494931538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,10240,0.01162577751610014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,65536,0.06113600068622165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,16384,0.04009777638647292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,16384,0.019646222392718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,2048,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,512,0.014387556248241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,6144,0.01869777838389079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,3584,0.006511111226346757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,768,0.03425777620739407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,512,0.005255110975768831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,512,0.013092444174819522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,12288,0.020089776979552377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,64,0.0035422220826148987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,256,0.01681777834892273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,7168,0.0232968893316057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,10240,0.05370222197638618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,65536,0.08380355437596639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,1024,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,768,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,256,0.005526222288608551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,512,0.011356444822417365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,3072,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,4096,0.00794666674402025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,128,0.10428444544474284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,3072,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,3584,0.007995555798212687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,10240,0.011764444410800934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,6144,0.0749066670735677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,6144,0.00904444439543618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,5120,0.00833422193924586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,512,0.003522666792074839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,8192,0.02416177756256527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,4096,0.013605333036846586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,2048,0.0052168890833854675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,8192,0.10301066769493951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,768,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,10240,0.01200799975130293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,12288,0.013399110900031196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,3072,0.05258488986227247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,1536,0.008258666429254744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,2048,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,16384,1.2681208716498482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,3072,0.020459555917316012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,256,0.08362044228447808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,128,0.019648000597953796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,768,0.006903111106819577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,2560,0.011996444728639392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,1024,0.010866666833559671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,64,0.007710222568776872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,12288,0.020785777105225455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,8192,0.013791110780504016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,512,0.031210667557186548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,32,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,128,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,1024,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,512,0.013070222404268054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,1024,0.0038648889296584656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,65536,0.15425066153208414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,3072,0.010747555229398938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,4096,0.040066666073269315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,16384,0.01612355477280087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,256,0.058598220348358154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,768,0.01753422286775377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,4096,0.00832088871134652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,1024,0.006439111298984951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,32,0.006451555424266391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,3584,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,65536,0.09054399861229791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,16384,0.1521244446436564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,8192,0.013274666335847644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,3072,0.007385777930418651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,3072,0.007677333222495184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,768,0.00703288863102595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,512,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,3584,0.00740622232357661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,3584,0.03502666619088914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,1536,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,256,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,768,0.005877333382765452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,10240,0.06642311149173312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,7168,0.04281066523657905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,4096,0.026683555708991155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,5120,0.00868000007337994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,2560,0.08633510933982001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,16384,0.017864000466134813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,12288,0.027871999475691054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,768,0.008961777720186446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,128,0.004939555707905027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,768,0.004506666627195147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,6144,0.03623466690381368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,2560,0.020224889119466145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,4096,0.016880000631014507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,512,0.010384889112578498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,3072,0.009725333087974125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,65536,0.12147200107574463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,5120,0.015571556157535978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,128,0.005374222166008419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,8192,0.01739644507567088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,1024,0.012288000020715924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,12288,0.050308446089426674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,1536,0.011759999725553723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,3072,0.007684444387753804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,2048,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,4096,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,12288,0.02528800070285797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,6144,0.1369697782728407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,7168,0.049601776732338794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,12288,0.014536889062987434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,5120,0.010241777532630498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,3072,0.027035555905765955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,65536,1.2242320378621419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,2560,0.04335644510057238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,512,0.005515555540720622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,1536,0.00869599978129069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,512,0.003407111184464561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,7168,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,256,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,65536,0.0695377786954244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,256,0.008642666869693333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,5120,0.008794666992293464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,64,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,1024,0.0038622220357259116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,1536,0.05723466475804647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,1024,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,128,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,7168,0.009711110757456886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,65536,0.03594844539960226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,8192,0.06577955351935492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,8192,0.1275751060909695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,8192,0.014141332772043018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,2560,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,2560,0.007250666618347168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,256,0.008028444316652086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,5120,0.008987555901209513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,3072,0.025085333320829604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,8192,0.015178667174445258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,3584,0.006583999842405319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,7168,0.08411733309427898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,16384,0.02131822208563487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,4096,0.04305422306060791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,1024,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,768,0.003559111307064692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,16384,0.033971554703182645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,16384,0.20447199874454072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,1536,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,8192,0.010752889017264048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,1024,0.019342222147517733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,3584,0.07682844665315416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,512,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,3072,0.011582222249772815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,64,0.007995555798212687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,6144,0.019468444916937087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,256,0.01718577742576599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,1536,0.018548443913459778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,3584,0.015267555912335714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,7168,0.02102311121092902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,2048,0.0069217777086628815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,7168,0.03607822126812405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,256,0.014144889182514615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,7168,0.021371554997232225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,4096,0.009555555880069733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,5120,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,3584,0.0075697774688402815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,3072,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,7168,0.019558222757445443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,3072,0.013481777575280933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,256,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,10240,0.017274666163656447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,2560,0.052560001611709595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,256,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,1024,0.003842666745185852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,4096,0.10011555751164754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,1536,0.010629333555698395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,512,0.015477332803938123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,3072,0.018831110662884183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,6144,0.14507910940382215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,6144,0.008972444468074376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,2560,0.01721600029203627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,7168,0.029890666405359905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,12288,0.07109333409203424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,16384,0.04203733470704821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,10240,0.0120319997270902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,2048,0.031011554929945204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,768,0.01995644470055898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,768,0.015007111761305066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,1024,0.017120000388887193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,128,0.012847111456924014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,3072,0.013367111484209696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,4096,0.020450666546821594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,16384,0.028839998775058325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,16384,0.6511528756883409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,10240,0.10786488983366226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,4096,0.011384000380833944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,10240,0.054633776346842446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,7168,0.014765333798196582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,2048,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,16384,0.2356719970703125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,6144,0.27789778179592556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,3072,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,512,0.006194666855865055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,2560,0.027073777384228174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,6144,0.12010578314463298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,6144,0.009733333355850643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,128,0.003867555409669876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,5120,0.012969777815871768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,12288,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,1024,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,768,0.013739555246300168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,2048,0.006592000110281839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,65536,0.2774657673305935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,3584,0.011507555842399597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,8192,0.02684088879161411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,12288,0.013179555535316467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,128,0.010694222317801582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,16384,0.31084977255927193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,8192,0.10612799723943074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,5120,0.00922400007645289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,6144,0.019295111298561096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,32,0.002833777831660377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,10240,0.019926221834288705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,2048,0.019395555059115093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,3584,0.01630311045381758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,3584,0.016543999314308167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,1024,0.004552000098758274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,12288,0.033682667546802096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,32,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,65536,0.16538488864898682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,2048,0.006528888725572162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,10240,0.024060444699393377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,7168,0.010031110710567897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,1536,0.007245333658324347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,3072,0.014489778214030795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,5120,0.017516444126764934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,256,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,7168,0.02053244411945343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,1536,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,8192,0.02196888956758711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,3072,0.012797333300113678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,8192,0.010770666930410596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,4096,0.014058666096793281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,1024,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,2048,0.05386133326424492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,2048,0.013226666384273104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,2048,0.008656000097592672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,10240,0.04018222292264303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,6144,0.1656737724939982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,768,0.01682577696111467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,1536,0.03219377663400438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,4096,0.24712888399759927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,10240,0.01329422245422999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,16384,0.08110400040944417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,5120,0.12117777930365668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,6144,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,4096,0.024404444628291663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,128,0.04346311092376709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,512,0.016851555969980028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,3584,0.014464888307783337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,65536,0.15955644183688694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,4096,0.01959466603067186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,16384,0.040962666273117065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,5120,0.02165599995189243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,64,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,65536,0.9151235156589084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,10240,0.18449600537618002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,2560,0.005738666488064661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,10240,0.02549422283967336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,512,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,2048,0.010301333334710863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,512,0.01828799976242913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,16384,0.02920444475279914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,8192,0.023911111884646948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,768,0.0225600004196167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,8192,0.019898666275872123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,65536,0.2943884531656901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,7168,0.16074577967325845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,6144,0.012089778151777057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,3584,0.01827022267712487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,3584,0.34380798869662815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,4096,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,4096,0.0123848890264829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,8192,0.016161777906947665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,16384,0.021186666356192693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,256,0.016515556308958266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,3072,0.014069333672523499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,6144,0.008719999757077958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,32,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,2560,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,64,0.0025111111915773815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,32,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,16384,0.015135111080275642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,65536,0.01995644470055898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,2560,0.014961777461899651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,32,0.0031946665710873077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,1536,0.004619555754794015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,5120,0.29254399405585396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,2048,0.014711999230914645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,16384,0.015770667129092746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,5120,0.022042666872342426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,7168,0.019691555036438834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,6144,0.021983999345037673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,3584,0.021817778547604878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,16384,0.018735999862353008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,8192,0.014678221609857349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,2048,0.031182222896152075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,10240,0.1991057793299357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,3584,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,65536,0.04930399854977926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,512,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,256,0.005555555638339784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,2048,0.010379555324713389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,3584,0.02201866606871287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,3072,0.011063111325105032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,512,0.006239999913507038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,5120,0.021161778105629817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,6144,0.02403733299838172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,1024,0.004229333251714706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,512,0.03410933415095011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,16384,0.024148444334665935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,65536,0.2483422226376004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,1536,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,5120,0.009329777624871995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,4096,0.09562489059236313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,4096,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,8192,0.03774311145146688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,512,0.023063111636373732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,65536,0.0658151110013326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,4096,0.015976889265908133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,512,0.0039048890272776284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,32,0.0038737778862317405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,2048,0.011720000041855706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,12288,0.02802844511138068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,2048,0.025063999825053748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,16384,0.06696889135572645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,65536,1.1614400015936959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,65536,0.2590577867296007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,4096,0.039312889178593956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,2048,0.014815110299322339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,256,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,512,0.0035093331502543557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,4096,0.009029332962301042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,1536,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,16384,0.012947555217478009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,2048,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,6144,0.01074577785200543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,128,0.015103111664454142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,768,0.07866311073303223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,12288,0.1825502183702257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,4096,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,8192,0.010339555641015371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,16384,0.02126666737927331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,8192,0.015226667126019796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,3584,0.021316443880399067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,5120,0.019726221760114033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,5120,0.016179554992251925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,768,0.010545777777830759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,5120,0.13124622239006892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,1536,0.016511999898486666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,8192,0.015114666687117683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,64,0.0032008888406885993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,1536,0.010647999743620554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,16384,0.2772880130343967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,2048,0.0169493340783649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,256,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,128,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,3584,0.008181333541870117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,8192,0.010830221904648675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,128,0.00544088871942626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,128,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,1024,0.012736000120639801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,6144,0.00775733341773351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,2560,0.006084444622198741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,3072,0.006745777610275481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,1536,0.016144000821643405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,2560,0.0070266665683852295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,2048,0.033612443341149226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,7168,0.2264151043362088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,5120,0.016114667057991028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,16384,0.028015110227796767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,6144,0.0666195551554362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,10240,0.028714666763941448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,65536,0.15317955281999376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,1536,0.009425777528021071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,3584,0.006304889089531369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,512,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,7168,0.03455911080042521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,6144,0.11520089043511285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,768,0.02050755586889055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,768,0.00793422261873881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,65536,0.06635822190178765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,768,0.020438222421540152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,5120,0.011724444727102915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,512,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,512,0.07732977469762166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,16384,0.20564711093902588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,5120,0.009722666607962714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,128,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,7168,0.018239999810854595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,7168,0.014702222413486905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,3584,0.0064195555945237475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,64,0.006311111152172089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,6144,0.014800889624489678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,32,0.002495999965402815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,3584,0.027837332752015855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,768,0.005219555563396878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,512,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,512,0.0052595556610160405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,7168,0.011511999699804517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,64,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,16384,0.01906488835811615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,4096,0.0076453329788313965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,4096,0.00794400026400884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,7168,0.01870577699608273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,3072,0.007687110867765214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,5120,0.009336888790130615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,7168,0.023013333479563396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,65536,0.060214221477508545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,65536,0.09700977802276611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,128,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,65536,0.07188089026345147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,16384,0.012055111428101858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,768,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,256,0.005219555563396878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,2048,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,512,0.0035208890007601846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,768,0.01648977730009291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,2048,0.02199466692076789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,2048,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,3584,0.019931554794311523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,8192,0.014723555909262763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,1536,0.004494222087992562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,3584,0.01204711116022534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,3584,0.013678222066826291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,6144,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,8192,0.010729778144094678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,3584,0.025322667426533167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,768,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,768,0.0064275554484791225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,8192,0.010687111152542962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,1024,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,12288,0.02732533382044898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,10240,0.011739555332395764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,3584,0.034033778640959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,12288,0.03327288892534044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,3584,0.010039110978444418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,256,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,3072,0.057337780793507896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,16384,0.017887110511461895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,512,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,16384,0.05736266904407077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,768,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,2048,0.006970666348934174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,65536,0.16291111045413548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,2560,0.007567110988828871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,768,0.019898666275872123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,3072,0.009389333426952362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,768,0.013442666994200813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,32,0.011680000358157687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,8192,0.014119999276267158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,8192,0.04926310976346334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,128,0.049545778168572314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,8192,0.006619555668698416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,64,0.020024889045291476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,512,0.004905777672926585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,8192,0.0237137774626414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,4096,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,3072,0.03510399990611606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,768,0.023796444137891132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,16384,0.07384800248675875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,65536,0.017906667457686532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,2560,0.006620444357395172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,65536,0.07646577888064914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,5120,0.1315573321448432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,5120,0.010621333287821876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,6144,0.02602044410175747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,16384,0.14590488539801702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,6144,0.017906667457686532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,512,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,65536,0.060784889592064746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,10240,0.012103111379676394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,2048,0.030165334542592365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,3072,0.010040000081062317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,64,0.011736888852384357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,6144,0.020245333512624104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,4096,0.017200888858901132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,1536,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,16384,0.020181333025296528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,65536,0.35254666540357804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,8192,0.011086222198274402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,5120,0.016314667132165697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,256,0.002941333378354708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,32,0.0028168888141711554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,4096,0.007992888490358988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,16384,0.1948915587531196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,32,0.0032035555276605818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,3072,0.08007111152013142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,3584,0.017294221454196505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,1024,0.015156444576051501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,5120,0.015602666470739575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,5120,0.009415999882751042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,4096,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,5120,0.008979555633332994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,3072,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,3072,0.009656888743241629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,5120,0.011257777611414591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,3584,0.014864888456132678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,3072,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,12288,0.01330311099688212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,1536,0.016116445263226826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,16384,0.0487937761677636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,1024,0.004202666795916027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,10240,0.012730666332774691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,3072,0.1406595574484931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,16384,0.05415999889373779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,256,0.008295111358165741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,3072,0.04842844605445862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,2048,0.006934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,2048,0.29841245545281303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,512,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,3072,0.009187555147541894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,7168,0.014495111174053617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,3072,0.016549333930015564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,7168,0.07722222142749362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,768,0.005571555760171678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,8192,0.015287111202875773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,2048,0.0458222230275472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,3584,0.04365422328313192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,512,0.014783110883500842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,6144,0.13644088639153376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,10240,0.03916800022125244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,2560,0.009685333404276106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,16384,0.031192888816197712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,3584,0.2300879955291748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,3584,0.01443377799457974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,7168,0.013744889034165276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,7168,0.019606222709019978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,256,0.006223111102978389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,65536,0.030352890491485596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,7168,0.20303555329640707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,3072,0.020213334096802604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,10240,0.014752000570297241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,256,0.0035004446076022256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,128,0.014070222775141398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,3072,0.009956444303194681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,128,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,2560,0.008434666527642144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,256,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,32,0.003216888962520493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,65536,0.1826417843500773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,7168,0.07641600237952338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,4096,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,3584,0.08237866560618083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,8192,0.019596444235907663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,12288,0.021128000484572515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,12288,0.017656889226701524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,6144,0.009071110851234859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,3584,0.007667555577225155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,16384,0.015503999259736804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,12288,0.020061333974202473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,5120,0.008621333373917473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,512,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,65536,0.4329377810160319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,7168,0.009702222214804756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,32,0.0031991110493739447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,16384,0.21753956211937797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,512,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,1024,0.0041582224269707995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,3584,0.019708444674809773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,3072,0.049525333775414355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,2560,0.007699555820888943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,256,0.017302221722073026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,64,0.006619555668698416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,32,0.0028479999552170434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,768,0.014485332700941296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,1536,0.007989333735571967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,16384,0.48344177669949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,2048,0.0743395553694831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,16384,0.10711733500162761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,16384,0.03602844476699829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,8192,0.028540445698632136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,5120,0.02914488977856106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,10240,0.05831199884414673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,1024,0.004898666507667965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,7168,0.14110667175716823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,7168,0.012664000193277994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,6144,0.014840889308187695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,10240,0.01651022169325087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,3072,0.010631110933091907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,32,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,1024,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,1024,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,768,0.0038604442444112566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,65536,1.9554125467936199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,128,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,128,0.01484444406297472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,1024,0.0038337777886125776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,16384,0.04659555686844719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,2560,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,128,0.010294222169452244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,256,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,128,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,12288,0.029488888051774766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,6144,0.0328764451874627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,6144,0.06778844197591145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,65536,0.08149600028991699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,1024,0.005223999834722943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,7168,0.00959644466638565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,768,0.007339555356237624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,128,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,768,0.024830222129821777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,6144,0.08316977818806966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,65536,0.05532177620463901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,2560,0.006737777756320105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,4096,0.02321333355373806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,1536,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,7168,0.047172443734274976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,512,0.02013066742155287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,768,0.006238222122192383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,10240,0.025610667135980394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,512,0.024886222349272832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,3584,0.010039110978444418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,12288,0.1503937774234348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,6144,0.10593866639667088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,65536,0.06630044513278537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,256,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,256,0.0032186667538351486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,8192,0.015305777390797934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,256,0.01651822196112739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,65536,0.01752622259987725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,512,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,12288,0.11536978350745307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,256,0.004218666503826777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,8192,0.01994400057527754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,1024,0.06824266910552979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,4096,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,512,0.007327111230956183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,3072,0.0063777777055899305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,12288,0.036822223001056247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,10240,0.3131973213619656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,16384,0.021194666624069214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,8192,0.06486399968465169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,7168,0.016020443704393175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,65536,0.23339909977383086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,768,0.01757866640885671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,8192,0.016335111525323655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,6144,0.01813155578242408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,65536,0.08782755666308933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,1536,0.0696586635377672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,768,0.0765520003106859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,1536,0.22669778929816353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,8192,0.03685422076119317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,16384,0.03988088832961188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,128,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,2048,0.009642666412724389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,256,0.01238844460911221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,128,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,12288,0.02271466619438595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,3584,0.008192000289758047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,16384,0.06289155615700616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,10240,0.020052444603708055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,16384,0.23787644174363878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,256,0.00286666676402092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,128,0.011387555963463254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,1536,0.015172445111804538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,6144,0.024626667300860088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,1536,0.004829333474238713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,128,0.003395555540919304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,64,0.007095110913117726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,3584,0.007789333661397298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,2560,0.0074240002367231585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,2560,0.013436444103717804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,10240,0.020558221472634208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,5120,0.012169777519173093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,32,0.008458666503429413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,4096,0.09817955229017471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,12288,0.010464000205198923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,1536,0.006351111249791251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,768,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,768,0.022784888744354248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,64,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,1536,0.004920888692140579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,768,0.003548444559176763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,10240,0.012082666986518435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,2048,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,256,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,5120,0.016892444756295946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,6144,0.024847111768192712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,4096,0.009018667042255402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,256,0.006636444479227066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,4096,0.08376622200012207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,5120,0.04338399900330437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,12288,0.01331022216214074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,3072,0.006326222171386083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,2560,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,256,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,12288,0.14111110899183485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,7168,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,1024,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,1024,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,128,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,32,0.016725333200560678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,2560,0.007287999822033777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,3072,0.06962311267852783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,512,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,65536,0.05513066715664334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,4096,0.01658933361371358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,1024,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,16384,0.03902488946914673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,8192,0.01072622256146537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,512,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,4096,0.017027555240525138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,768,0.014481777946154276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,8192,0.08449955781300862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,6144,0.020201777418454487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,2560,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,3584,0.009456000394291347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,1024,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,64,0.0033475555893447665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,3072,0.013771555489963956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,256,0.00351555562681622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,512,0.17797599898444283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,65536,0.6889591217041016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,2048,0.018226666582955253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,7168,0.02875022093454997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,4096,0.011938666303952536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,5120,0.012440889245933957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,16384,0.023377777801619634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,5120,0.01815555493036906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,2048,0.015140444040298462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,12288,0.14783999654981825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,10240,0.04973777797487047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,6144,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,512,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,128,0.0038480001191298165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,2560,0.03181688984235128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,5120,0.01750577820671929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,768,0.006238222122192383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,4096,0.009611555271678502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,128,0.0031991110493739447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,5120,0.018211555149820115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,768,0.017873777283562552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,8192,0.017383999294704862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,2560,0.009396444592210982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,128,0.015506666567590503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,1024,0.025771554973390367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,4096,0.029481778542200725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,16384,0.05742488967047798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,3072,0.010966222319338055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,16384,0.07972444428337945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,16384,0.05782844622929891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,8192,0.27850400076972115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,3072,0.009304888546466827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,16384,0.020266667008399963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,16384,0.023623999622133043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,512,0.003825777934657203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,3584,0.016562667157914903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,512,0.0038497779104444715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,6144,0.021715554926130507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,3584,0.011077333655622272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,7168,0.012630222572220696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,65536,0.07691111167271932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,1536,0.0701191094186571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,256,0.003474666840500302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,7168,0.019741333193249173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,5120,0.05724177757898966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,128,0.012086222569147745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,3584,0.02902844548225403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,3072,0.006567111031876669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,10240,0.03360177742110358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,16384,0.13574488957722983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,768,0.006945778098371293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,10240,0.011432888607184092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,7168,0.02255200015174018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,10240,0.09571378098593818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,3072,0.008026666939258575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,1024,0.035487999518712364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,5120,0.008945778012275696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,3584,0.016971555021074083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,3584,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,16384,0.026058667235904272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,128,0.005525333185990651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,128,0.005542222410440445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,3584,0.013622221847375235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,128,0.015434667468070984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,7168,0.06276000208324857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,128,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,3072,0.00683822234471639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,1024,0.019810666640599568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,4096,0.01627288924323188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,6144,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,65536,0.1646302276187473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,1536,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,3584,0.05949778027004666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,1024,0.02439466615517934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,12288,0.017216889394654166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,2048,0.020296888218985665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,4096,0.14825688468085393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,128,0.01818933255142636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,7168,0.039566222164365984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,5120,0.025921778546439275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,2560,0.02328444520632426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,10240,0.03492088781462775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,10240,0.029271112547980413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,10240,0.023772444989946153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,3072,0.02880355715751648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,6144,0.02068177858988444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,128,0.0052426668504873914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,32,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,768,0.0061919999619325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,8192,0.013243555194801755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,2560,0.007962666451931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,5120,0.01921777758333418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,1536,0.0068417779273457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,1536,0.01891644464598762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,7168,0.009700444837411245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,2560,0.01254755589697096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,512,0.019111999207072787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,512,0.014124444789356656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,4096,0.37506667772928876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,768,0.004934222333961063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,256,0.00564177789621883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,2560,0.012792888614866467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,1536,0.04871999886300829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,768,0.021700445148679946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,2048,0.013491555220550962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,2560,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,8192,0.015136889285511441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,768,0.004541333350870344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,1536,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,3072,0.015104888214005364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,128,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,6144,0.02917511264483134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,65536,0.0621582203441196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,10240,0.0437084436416626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,128,0.00553688903649648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,2560,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,1536,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,6144,0.009487110707494948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,128,0.013771555489963956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,256,0.003175111073586676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,3072,0.015810666812790763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,768,0.00904888908068339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,2048,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,4096,0.09509599871105617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,128,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,1536,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,1024,0.026242666774325903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,768,0.01514488955338796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,768,0.0035537779331207275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,1536,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,6144,0.009660444325870937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,768,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,12288,0.03042666779624091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,3072,0.17350933286878797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,3072,0.012390221986505719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,65536,0.15552177694108751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,6144,0.058787557813856334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,12288,0.028363555669784546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,4096,0.011354666617181567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,256,0.0035208890007601846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,256,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,3584,0.0069164443347189165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,1024,0.021309332715140447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,7168,0.12467200226253933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,3072,0.0074551113777690465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,512,0.013600889179441663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,8192,0.010692444112565784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,16384,0.016162667009565566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,16384,0.021029333273569744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,6144,0.024780443973011438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,512,0.02219999995496538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,5120,0.386353784137302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,5120,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,65536,0.08647733264499241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,6144,0.048595554298824735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,512,0.003822222352027893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,7168,0.013411555853154926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,6144,0.09442400270038181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,768,0.006367110957702001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,6144,0.013190222283204397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,2048,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,2560,0.005816000203291575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,1024,0.003879111260175705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,32,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,5120,0.00868088917599784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,256,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,2048,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,1536,0.014158222410413953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,10240,0.027320888307359483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,1024,0.00453333349691497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,32,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,7168,0.09856533341937596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,10240,0.018183110488785636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,5120,0.01757244434621599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,8192,0.025781333446502686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,5120,0.01686844395266639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,65536,0.03038311004638672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,2048,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,3584,0.007972444097201029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,8192,0.021738666627142165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,1536,0.019253333409627277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,65536,0.07092444764243232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,768,0.03490044342146979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,7168,1.963358137342665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,3072,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,2560,0.026800000005298193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,12288,0.05205511053403219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,10240,0.02123111155298021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,1024,0.007691555553012424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,1024,0.02875111169285244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,12288,0.07756621970070733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,512,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,128,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,3584,0.009072889056470659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,256,0.015159999330838522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,768,0.014900444282425774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,64,0.0028106667515304354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,256,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,10240,0.4015066623687744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,4096,0.009395555489593083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,5120,0.0945804450247023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,3072,0.04948088857862684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,6144,0.018904000520706177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,3072,0.007631999750932057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,768,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,6144,3.396343019273546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,3584,0.027423111928833857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,2048,0.013396444420019785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,768,0.004232888834344016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,12288,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,768,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,768,0.005640000104904175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,10240,0.02510044475396474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,128,0.010674667027261523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,12288,0.0561386677953932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,3584,0.009416888985368941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,2048,0.027130666706297133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,8192,0.01071911139620675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,3584,0.03228444523281521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,2560,0.008252444366614023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,6144,0.02348533272743225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,3072,0.007449777589903937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,4096,0.01934400035275353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,64,0.010393777655230628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,6144,0.034190220965279475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,1536,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,2560,0.0058613332609335584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,3584,0.007599111232492659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,12288,0.009717333647939894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,7168,0.03692266676161025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,768,0.11552622583177354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,32,0.0032053333189752367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,256,0.009359111388524374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,3072,0.042071110672420926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,64,0.006027555714050929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,5120,0.01887999971707662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,5120,0.008960888617568547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,256,0.0028799999919202593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,16384,0.022317333353890315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,128,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,16384,0.06580266687605116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,65536,0.16392444239722356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,3584,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,16384,0.031850665807724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,7168,0.03672800130314297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,12288,0.03084711233774821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,7168,0.017088888419999015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,512,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,6144,0.02379377848572201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,6144,0.048039999273088246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,4096,0.011057777537239922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,128,0.011140444212489657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,3072,0.012464889221721225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,12288,0.013110222087966071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,16384,0.0977404448721144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,3584,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,1024,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,128,0.024979554944568213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,65536,0.12843288315667048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,512,0.0041955556306574075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,6144,0.012679110798570844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,128,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,2560,0.015637333194414776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,128,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,2048,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,512,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,10240,0.035042666726642184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,3072,0.012335110869672565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,3072,0.009530666801664565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,4096,0.02106933295726776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,4096,0.016235555211702984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,6144,0.018529777725537617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,2560,0.07224444548288982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,4096,0.011769777370823754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,16384,0.04794311192300585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,256,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,1024,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,3072,0.007213333414660559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,1536,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,16384,0.06500177913241915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,1536,0.03695822093221877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,1024,0.004889777965015835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,1536,0.012799999780125089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,4096,0.09503733449512058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,10240,0.02538577715555827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,512,0.015465777781274585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,65536,0.03868888815244039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,2048,0.016887111796273124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,2560,0.011072888970375061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,768,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,3072,0.03713333275583055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,4096,0.009434666898515489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,6144,0.009684444301658207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,768,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,256,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,16384,0.1682737800810072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,32,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,65536,0.05711999866697523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,6144,0.018937778141763475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,256,0.012692444854312472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,1536,0.019893333315849304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,1536,0.09923822349972194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,768,0.003674666500753827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,2560,0.08937511179182266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,512,0.0035128887328836652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,128,0.01514311134815216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,3584,0.01519466605451372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,2560,0.007258666886223688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,128,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,10240,0.02833777666091919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,256,0.005417777846256892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,3584,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,12288,0.013086222112178802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,3072,0.02367466688156128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,5120,0.03618755605485704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,5120,0.01145155562294854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,3072,0.013878222141000958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,1024,0.03728355632887946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,32,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,7168,0.009658666948477427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,12288,0.018926221463415358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,32,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,65536,0.07019466824001737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,16384,0.07124622000588311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,16384,0.009668444593747457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,6144,0.03601866629388597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,6144,0.018560889694425795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,10240,0.02536533276240031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,65536,0.035135110219319664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,12288,0.10276089111963908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,3072,0.01480888823668162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,16384,0.17146666844685873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,2048,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,6144,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,12288,0.018269333574506972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,5120,0.012440000143316058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,256,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,10240,0.11886044343312581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,5120,0.02422399984465705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,768,0.015022221538755627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,12288,0.24411466386583117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,8192,0.01091733326514562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,65536,0.05350044369697571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,128,0.0028719999310043124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,128,0.005222222043408289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,7168,0.010205333431561789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,768,0.006186666587988536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,128,0.008143111235565608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,6144,0.009479111267460717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,4096,0.040629333919949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,1024,0.011354666617181567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,4096,0.018907555275493197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,128,0.014851555228233337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,65536,0.07557333177990384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,12288,0.018248889181349013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,128,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,64,0.0030213333666324615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,512,0.003547555456558863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,5120,0.0403386652469635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,768,0.011371555427710215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,16384,0.015928000211715698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,5120,0.09454844395319621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,5120,0.009476444787449306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,3072,0.0074622225430276655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,256,0.016984888248973422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,5120,0.018799111247062683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,256,0.0036151111125946045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,768,0.0074879998962084455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,768,0.008052444292439355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,1536,0.05055377880732218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,10240,0.02437688906987508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,6144,0.09704355398813884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,1536,0.006215999937719769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,65536,0.046626667181650795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,3584,0.2233804331885444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,128,0.003910222401221593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,12288,0.015108444624476962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,2048,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,65536,0.20073955588870576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,768,0.03629688753022088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,7168,0.02204977803760105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,16384,0.036885334385765925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,10240,0.14605066511366102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,256,0.0052391112678580815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,768,0.016177778442700703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,2048,0.023520888553725347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,1536,0.00674222202764617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,1536,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,3072,0.5746089087592231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,3584,0.05098933312628004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,7168,0.012248000337017907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,6144,0.01789866718981001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,512,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,10240,0.10950311024983723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,2048,0.011759999725553723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,12288,0.008571555217107138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,512,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,65536,0.3008444574144152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,12288,0.014110222458839417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,32,0.01789777808719211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,3584,0.015855110353893705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,3584,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,3584,0.007907555335097844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,2048,0.016156444946924847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,5120,0.009024000002278222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,16384,0.027895109521018133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,128,0.005534222142563925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,768,0.005636444522274866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,6144,0.010370666782061258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,64,0.0028524444335036804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,3584,0.03222399950027466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,512,0.017533333765135873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,1536,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,10240,0.015249777171346875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,64,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,128,0.0028168888141711554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,16384,0.01570755574438307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,16384,0.13399910926818848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,128,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,64,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,128,0.013062222136391534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,256,0.01662133302953508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,1536,0.006215111249023014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,6144,0.058485335773891874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,3584,0.0233262214395735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,128,0.01512977812025282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,8192,0.01961955593691932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,7168,0.010417777631017896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,2048,0.011058666639857821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,2560,0.006385777973466449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,6144,0.0355422231886122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,65536,0.05899644560284085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,64,0.030635555585225422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,128,0.0031315556002987754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,7168,0.013437333206335703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,6144,0.018219555417696636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,16384,0.025460445218616064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,1024,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,12288,0.04411822226312426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,2560,0.00963644435008367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,256,0.013190222283204397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,256,0.006520888871616787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,12288,0.030396444929970637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,65536,0.1417191161049737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,512,0.01205244412024816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,1536,0.06741066773732503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,3584,0.017512889371977914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,128,0.005394666559166378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,65536,0.05787555376688639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,8192,0.014856000741322836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,65536,5.373138851589626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,4096,0.017553778158293832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,2560,0.01071111112833023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,7168,0.023006222314304773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,10240,0.04191733400026957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,5120,0.0810497800509135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,64,0.0028799999919202593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,512,0.0058808889653947614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,8192,0.022945778237448797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,16384,0.027107555005285475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,65536,4.873945871988933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,65536,0.06499377886454265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,12288,0.31247467464870876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,6144,0.18076444996727836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,1024,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,12288,0.041951109965642296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,4096,0.020728000336223178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,1024,0.014774221513006421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,2048,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,7168,0.02126755482620663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,6144,0.009662222531106737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,7168,0.01850400037235684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,7168,0.06373955806096394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,512,0.003198222153716617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,6144,0.022797332869635686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,65536,0.020960888928837244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,256,0.015127110812399121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,1024,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,1024,0.01329422245422999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,8192,0.08052000072267321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,12288,0.03019022279315525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,768,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,2048,0.015807999504937064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,65536,0.11893155839708115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,512,0.01940444442960951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,3072,0.024373332659403484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,64,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,10240,0.025847999585999384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,4096,0.009352000223265754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,16384,0.044144001271989614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,7168,0.010020444790522257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,10240,0.030268443955315485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,5120,0.012168889244397482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,5120,0.012720000412729053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,128,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,12288,0.11741333537631565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,4096,0.07551911142137316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,6144,0.009298666483826106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,16384,0.035994665490256414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,16384,0.21705333391825357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,5120,0.00904088881280687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,6144,0.009128000173303816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,2560,0.012701333396964602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,3584,0.021516443954573736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,7168,0.010670222342014313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,2560,0.015822221835454304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,1536,0.014347554908858405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,2048,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,5120,0.01204533295498954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,128,0.01036444471942054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,2560,0.009191110730171204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,1024,0.028360890017615423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,128,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,16384,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,4096,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,6144,0.00942755573325687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,64,0.0028133332315418455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,6144,0.011075555450386472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,6144,0.2903670999738905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,3584,0.008643555144468943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,1024,0.03133777777353922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,512,0.004522666749027041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,3072,0.013830222189426422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,128,0.014793778459231058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,1024,0.018034666776657104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,3584,0.00870222184393141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,2048,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,3584,0.10163999928368463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,256,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,65536,0.09680355257458156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,768,0.016511110795868766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,7168,0.009666666388511658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,10240,0.025430222352345783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,2048,0.008644444247086843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,3072,0.016175111134847004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,2048,0.013804444836245643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,768,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,5120,0.02288533416059282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,768,0.006956444846259222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,7168,0.008995555341243744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,5120,0.018569777409235638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,128,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,7168,0.009333333207501305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,10240,0.024687111377716064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,512,0.0032720000793536506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,3072,0.027127999398443434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,1536,0.015821332732836407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,64,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,3072,0.00869777798652649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,5120,0.02150933278931512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,512,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,8192,0.013417777915795645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,3584,0.018031999468803406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,512,0.00351555562681622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,6144,0.008676444490750631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,5120,0.013344888885815939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,8192,0.030505776405334473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,1024,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,6144,0.04814399944411384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,7168,0.015095111396577625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,64,0.0032035555276605818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,12288,0.035535110367669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,64,0.011774222056070963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,65536,0.42756088574727374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,3072,0.007207999626795451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,768,0.003936888857020272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,2048,0.006639110959238476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,16384,0.04901599884033203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,10240,0.013271999855836233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,2560,0.015591111448076038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,3072,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,12288,0.05992622507943047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,128,0.007720888488822513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,7168,0.009778666827413771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,6144,0.029900444878472224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,3072,0.0075911109646161394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,1536,0.0083155557513237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,65536,0.06563466787338257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,65536,0.08900177478790283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,2048,0.009364444348547194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,8192,0.013432000246312885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,12288,0.02155911094612545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,2560,0.010759111079904767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,6144,0.013460444079505073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,10240,0.016155555844306946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,3584,0.017421333326233756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,512,0.0277839998404185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,3584,0.009707555174827576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,8192,0.02867022156715393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,7168,0.037541333172056414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,32,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,64,0.009679111341635386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,3072,0.01481244464715322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,256,0.01789066692193349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,256,0.0031795555518733132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,16384,0.036201778385374285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,768,0.006541333264774746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,7168,0.018535999788178336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,256,0.005192888693677055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,5120,0.016570667425791424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,2560,0.007278222176763747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,7168,0.021968000464969214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,512,0.015843555331230164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,8192,0.025086222423447505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,65536,0.08175822099049886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,512,0.0032337777730491427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,4096,0.09638133313920762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,12288,0.03489866521623399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,4096,0.03707022137112088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,4096,0.02377333409256405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,768,0.003510222252872255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,16384,0.08239466614193387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,3584,0.015856888559129503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,16384,3.0343040890163846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,2560,0.014816888504558139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,4096,0.015563555889659457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,512,0.016179554992251925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,8192,0.022310222188631695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,256,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,256,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,512,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,5120,0.025063110722435847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,2560,0.016248888439602323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,256,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,6144,0.012984889249006907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,1536,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,65536,1.9245537651909723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,8192,0.02092355489730835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,8192,0.022847111026446026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,3072,0.006198222024573221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,128,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,5120,0.009008000294367472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,256,0.006597333484225803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,128,0.0069217777086628815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,8192,0.012666666673289405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,16384,0.07635733154084948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,16384,0.031834665271970965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,2048,0.030792888667848375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,64,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,32,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,3072,0.007267555428875818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,2048,0.11825422445933025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,65536,0.010781333678298526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,128,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,16384,0.027658666173617046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,2560,0.023383999864260357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,4096,0.016598222984208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,2560,0.010454222559928894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,64,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,64,0.004886222382386525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,512,0.0038240001433425476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,8192,0.04826844400829739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,6144,0.028583112690183852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,1024,0.01349511080318027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,5120,0.04546666807598538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,768,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,3584,0.013812444276279874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,7168,0.14745510949028864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,12288,0.036666668123669095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,2560,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,65536,0.09797244601779515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,1536,0.004584889031118817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,32,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,12288,0.03573688864707947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,12288,0.014127110441525778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,6144,0.011935110721323224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,6144,0.010120888551076254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,8192,0.13939644230736628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,16384,0.022991999983787537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,32,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,10240,0.020920000142521326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,256,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,10240,0.011738667057620155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,6144,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,64,0.006222222414281633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,128,0.003132444495956103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,10240,0.38367022408379453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,512,0.00424799985355801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,3584,0.02757511039574941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,768,0.009386666946940953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,2560,0.012741333080662621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,2560,0.012601777911186218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,3584,0.012319111161761813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,768,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,5120,0.008264000217119852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,10240,0.09024533298280503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,2560,0.006950221955776215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,1536,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,32,0.0031431110368834603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,32,0.0035217776894569397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,3584,0.00886222223440806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,6144,0.012144000165992312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,8192,0.43066933419969344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,3072,0.008632889224423302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,12288,0.020256888535287645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,3584,0.011165333290894827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,4096,0.01514755520555708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,1024,0.004052444464630551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,256,0.004181333300140169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,128,0.006934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,32,0.0035377778112888336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,10240,0.027061333258946735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,512,0.053132444620132446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,3584,0.09247910976409912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,12288,0.027922666735119287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,6144,0.022256000174416438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,768,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,128,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,2560,0.011533333195580376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,32,0.0031066667288541794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,6144,0.11006577809651692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,7168,0.010058666268984476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,7168,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,2048,0.006771555377377405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,3584,0.020627554919984605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,3072,0.006263999889294307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,6144,0.01677600045998891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,10240,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,16384,0.024665777881940205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,4096,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,65536,0.1285760005315145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,3584,0.008332444561852349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,1024,0.004492444296677907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,512,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,6144,0.01204622205760744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,32,0.012736000120639801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,10240,0.02516355613867442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,8192,0.49051464928521055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,16384,0.04810400141610039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,2048,0.015590222345458137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,1536,0.027806222438812256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,65536,0.2433991167280409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,1024,0.006038222048017714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,64,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,2560,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,128,0.014539556370841132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,6144,0.016562667157914903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,1536,0.013802666631009845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,256,0.018222222725550335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,512,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,6144,0.01683022247420417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,10240,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,3072,0.02681333323319753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,3072,0.032024890184402466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,7168,0.009891555541091496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,256,0.02718222141265869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,12288,1.0587422053019206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,3072,0.018212444252438016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,3584,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,2560,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,12288,0.03382577829890781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,2048,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,6144,0.3017626603444417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,1024,0.017483555608325534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,32,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,2560,0.009333333207501305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,8192,0.01853511068556044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,512,0.007896000312434303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,16384,0.02644444505373637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,5120,5.745638953314887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,3584,0.011018666956159802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,16384,0.03657689028316074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,65536,0.09676000144746567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,3072,0.11874311500125462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,16384,0.0215022216240565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,10240,0.08809155225753784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,1024,0.020299555526839364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,3584,0.014446222119861178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,3584,0.014928888943460254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,5120,0.009313777916961247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,8192,0.010393777655230628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,1024,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,16384,0.04307111104329427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,2048,0.2264711062113444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,10240,0.013816888961527081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,256,0.00328533330725299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,5120,0.044904887676239014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,1536,0.012465777496496836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,3072,0.022502221994929846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,10240,0.037575999895731606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,512,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,768,0.4794613520304362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,7168,0.01721955504682329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,2560,0.028135998381508723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,1536,0.0196079992585712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,4096,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,512,0.0052915554907586836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,1024,0.014163555370436775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,10240,0.1994995541042752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,3584,0.006955555743641323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,3072,0.008838222258620793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,4096,0.03122044603029887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,1024,0.011928000383906893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,3584,0.016389333539538913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,512,0.0038453332251972626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,64,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,768,0.24393423398335776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,10240,0.009316444396972656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,3584,0.009221333596441481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,7168,0.021272889441914026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,1024,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,5120,0.008085333638721043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,768,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,2048,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,4096,0.021013332737816706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,3584,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,2048,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,10240,0.008976889153321585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,65536,0.06052977508968777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,4096,0.013389333254761167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,1536,0.016520000166363187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,128,0.00553955551650789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,7168,0.014116444521480136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,128,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,1024,0.0042364444169733255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,2048,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,5120,0.013481777575280933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,65536,0.023783110909991797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,512,0.006663110935025745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,5120,0.007986666427718269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,3072,0.010703110860453712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,128,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,16384,0.08663644393285115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,512,0.012792888614866467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,6144,0.1557111077838474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,4096,0.053080889913770885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,768,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,2048,0.006912888752089606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,3072,0.01071200023094813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,7168,0.01203644441233741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,256,0.015457777513398064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,5120,0.03465244505140517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,2048,0.014816888504558139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,65536,0.08181689182917277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,2048,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,256,0.0052533331844541765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,8192,0.022321777211295232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,7168,0.028026666906144884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,65536,0.11432266235351562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,8192,0.007260444263617198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,16384,0.04148088892300924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,1536,0.009709333380063375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,8192,0.006970666348934174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,3584,0.061215996742248535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,1024,0.014428445034556918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,4096,0.009764444496896531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,768,0.017817777064111497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,64,0.002536000063021978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,1024,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,1536,0.004909333255555895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,4096,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,2048,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,8192,0.07584710915883382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,6144,0.24468800756666395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,512,0.012741333080662621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,512,0.012418666647540199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,3584,0.011787555283970304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,6144,0.017690666847758822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,64,0.013068444199032254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,2048,0.005326222214433882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,16384,0.06432622008853488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,65536,0.062385777632395424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,128,0.00675733346078131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,65536,0.04703199863433838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,10240,0.2763813336690267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,12288,0.0307608875963423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,128,0.003206222214632564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,16384,0.02016266683737437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,768,0.03409511182043288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,12288,0.2986453374226888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,1536,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,4096,0.02292799949645996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,10240,0.02184533410602146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,7168,0.010835555692513784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,1024,0.007637333538797166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,5120,0.0507279998726315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,768,0.00424177779091729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,256,0.026900443765852187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,3584,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,64,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,6144,0.012469333079126147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,16384,0.025967111190160114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,5120,0.02141688929663764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,10240,0.02688177757793003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,3584,0.17255377769470215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,2560,0.01719111038578881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,2560,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,8192,0.3427297804090712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,3584,0.01351288871632682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,128,0.005210666606823604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,3072,0.05824177794986301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,5120,0.02084355552991231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,4096,0.02405422263675266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,512,0.022688888841205176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,1024,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,3072,0.008934222161769867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,7168,0.14167733987172446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,8192,0.00924355536699295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,5120,0.008346666892369589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,5120,0.029265777932273015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,6144,0.017954667409261067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,128,0.00923644420173433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,3072,0.009466666314336989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,6144,0.01683200067943997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,768,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,4096,0.02442488902144962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,1024,0.004591111093759537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,768,0.02233244478702545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,256,0.003136000078585413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,65536,0.31644267506069607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,1536,0.01754844519827101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,7168,0.019479999939600628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,256,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,5120,0.029439111550649006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,5120,0.020091555184788175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,2560,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,256,0.01687466601530711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,256,0.003208000005947219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,16384,0.020621332857343886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,2048,0.012089778151777057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,7168,0.010327111515733931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,1536,0.013108444710572561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,7168,0.009660444325870937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,768,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,65536,0.05890755520926582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,3072,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,8192,0.010667555862002902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,6144,0.03864977757136027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,512,0.006620444357395172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,1536,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,64,0.007652444144090016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,10240,0.01573066744539473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,128,0.005236444373925527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,64,0.011086222198274402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,8192,0.03227555420663621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,65536,0.01834755473666721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,10240,0.026591110560629103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,10240,0.013563555147912769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,2560,0.01552088889810774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,768,0.00792711145348019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,5120,0.24012178844875762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,12288,0.013798221945762634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,6144,0.021338666478792827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,2560,0.013593778014183044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,512,0.013077333569526672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,512,0.01368799971209632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,5120,0.007800000409285228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,2560,0.06427999999788073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,7168,0.009364444348547194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,6144,0.09137511253356934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,768,0.01717688971095615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,64,0.0038737778862317405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,3584,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,32,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,6144,0.009401777552233802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,12288,0.0676835576693217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,256,0.01311644415060679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,4096,0.08858666817347209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,4096,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,2048,0.009334222310119206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,2048,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,3584,0.04997511042488945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,8192,0.019909333851602342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,128,0.01481066644191742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,3584,0.013544888959990608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,512,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,3584,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,1024,0.03206844462288751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,768,0.0038933331767717996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,32,0.005984888722499211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,1024,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,64,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,128,0.014471999473041959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,2560,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,2048,0.01647644407219357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,3072,0.0070462218589252895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,32,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,10240,0.012552888856993781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,7168,0.07566310962041219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,256,0.002982222164670626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,32,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,2048,0.006894222150246303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,768,0.0038373333712418876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,4096,0.09965778059429592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,16384,0.03375466664632162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,3072,0.006924444602595435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,768,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,65536,0.07792800002627902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,6144,0.03038755390379164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,65536,0.16194844245910645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,4096,0.009071110851234859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,6144,0.015171556009186639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,3072,0.007606222397751278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,256,0.015153777268197803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,6144,0.009017777939637503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,256,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,4096,0.060644441180759005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,256,0.015200000670221118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,65536,0.18280888928307426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,5120,0.009492444495360056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,64,0.004883555488453971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,16384,0.011747555600272285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,7168,0.20594399505191377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,10240,0.009275555610656738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,10240,0.01955911186006334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,3072,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,3584,0.029343111647499934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,4096,0.00795555528667238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,1536,0.023022222850057814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,10240,0.024463999602529738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,6144,0.19152532683478463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,4096,0.01460355520248413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,16384,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,256,0.019283556275897555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,6144,0.025814221964942083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,512,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,3584,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,10240,0.02609511050913069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,512,0.01443911095460256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,3072,0.010729778144094678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,8192,0.02695466743575202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,1024,0.04831199844678243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,512,0.015485333071814643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,3072,0.01947022146648831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,128,0.005407111098368962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,1024,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,2048,0.07131199704276191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,10240,0.014821334017647637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,12288,0.031344887283113264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,10240,0.012476444244384766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,768,0.01552177800072564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,512,0.003925333420435588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,7168,0.007947555846638149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,768,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,1024,0.006219555520349079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,2560,0.007595555649863348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,768,0.05022933416896396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,32,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,2048,0.026309332913822595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,16384,0.039317333035998873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,7168,0.00960177762640847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,32,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,1536,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,2560,0.015528000063366361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,2048,0.0069004446268081665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,10240,0.015266666809717814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,6144,0.02956177790959676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,16384,0.016211556063758004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,3072,0.007128888534175024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,5120,0.010441777606805166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,512,0.006883555402358373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,1024,0.013739555246300168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,3072,0.0073831114504072405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,512,0.00388355553150177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,128,0.021645334031846788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,3584,0.05558311276965671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,3584,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,10240,0.021627555290857952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,10240,0.161624895201789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,512,0.007701333198282454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,1536,0.011578666667143503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,10240,0.014449778530332776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,8192,0.014918223023414612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,3072,0.00683111117945777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,3072,0.01683644453684489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,65536,0.04791288905673557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,16384,0.036116444402270846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,12288,1.9037395053439672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,2560,0.010324444207880232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,128,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,5120,0.01684266659948561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,64,0.0028239999794297745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,16384,0.02011644509103563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,65536,0.07828622394137912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,2048,0.006960888703664144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,2048,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,7168,0.12415911091698541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,7168,0.320287995868259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,2048,0.016913778252071805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,5120,0.007977777885066139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,1024,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,768,0.012249777714411417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,256,0.0031759999692440033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,8192,0.09527022308773464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,5120,0.008967111508051554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,512,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,512,0.0038693332009845306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,10240,0.019215111931165058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,1536,0.004523555437723796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,3072,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,7168,0.3157671027713352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,10240,0.036869333850012884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,256,0.004875555634498596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,3072,0.006989333364698622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,10240,0.025643555654419795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,256,0.007913777397738563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,16384,0.051310221354166664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,128,0.08343911170959473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,768,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,4096,0.009919111099508073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,16384,0.03156177865134345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,3584,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,10240,0.025581333372328017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,16384,0.016564443707466125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,128,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,512,0.003560888684458203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,64,0.008172444171375698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,64,0.0032604444358083936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,64,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,12288,0.15038400226169163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,6144,0.012698666916953193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,2560,0.00656444455186526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,128,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,8192,0.05968622366587321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,6144,0.013414222333166333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,12288,0.02676533328162299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,768,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,64,0.036999109718534685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,16384,0.03021866745418972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,256,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,32,0.00554577757914861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,5120,0.052525334888034396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,2048,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,6144,0.01826755536927117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,512,0.019780443774329293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,4096,0.007672888537247975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,16384,0.023444443941116333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,5120,0.012692444854312472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,5120,0.03177511029773288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,512,0.00328177772462368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,1024,0.005695111221737332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,64,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,3072,0.005660444498062134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,2560,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,65536,4.639588250054254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,8192,0.1143840021557278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,512,0.007055111229419708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,768,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,8192,0.006636444479227066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,6144,0.009705777797434065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,768,0.01720266706413693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,3072,0.007987555530336168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,64,0.0028435554769304064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,4096,0.008023111356629265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,2048,0.009481777747472128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,3584,0.01607644392384423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,64,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,65536,0.22692977057562935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,6144,0.027072888281610277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,16384,0.05632266733381483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,12288,1.546432918972439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,1024,0.005703111075692707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,7168,0.03087644444571601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,1536,0.006572444405820634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,8192,0.10604533221986558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,3072,0.01187111106183794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,16384,0.02957333458794488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,128,0.012680889003806643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,4096,0.011395555403497485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,65536,0.07134044170379639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,1536,0.006243555496136348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,2560,0.019310222731696237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,3072,0.007697777615653143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,1536,0.004646222210592694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,3584,0.07329955365922716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,2560,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,7168,0.010226666927337646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,1536,0.006305777778228124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,7168,0.034862223598692156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,2560,0.015830222103330824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,5120,0.021273778544531927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,3072,0.5608488718668619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,512,0.0052737775776121355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,512,0.01681777834892273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,64,0.006158222340875202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,4096,0.033518221643235944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,8192,0.09765777985254924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,16384,0.030994666947258845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,128,0.023374223046832617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,5120,0.012824888858530255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,65536,0.04624533322122362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,1536,0.014817777607176038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,16384,0.032015111711290144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,1024,0.02367199957370758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,16384,0.0255022214518653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,768,0.02276800076166789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,8192,0.010415111151006486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,1536,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,128,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,65536,0.0635875529713101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,3072,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,1536,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,10240,0.02475111186504364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,1536,0.03803022371398078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,12288,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,10240,0.014165333575672574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,12288,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,12288,0.02606755495071411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,3072,0.01515111161602868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,4096,0.015511999527613321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,128,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,10240,0.10392177767223781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,65536,0.07755111323462592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,32,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,10240,0.01036977767944336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,768,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,16384,0.08519111076990764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,3072,0.03177333209249709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,1024,0.008468444148699442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,256,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,2048,0.015840000576443143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,65536,0.6523066626654731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,2560,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,1536,0.04482577906714546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,3072,0.0075866662793689305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,2048,0.009355555805895064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,512,0.0069466663731469055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,2560,0.008071111308203803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,1024,0.0074471111098925276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,3072,0.02932888931698269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,256,0.005241777747869492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,128,0.0028364445186323593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,12288,0.18822133541107178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,12288,0.013432000246312885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,256,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,128,0.00656444455186526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,4096,0.007303111255168915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,7168,0.01827022267712487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,3072,0.01071466671095954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,128,0.017535110314687092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,32,0.0058808889653947614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,3072,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,768,0.006994666324721442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,3072,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,1024,0.014671110444598727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,512,0.01587288909488254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,128,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,256,0.003524444583389494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,3072,0.007293333609898885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,10240,0.011893333660231696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,1536,0.005439110928111606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,768,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,256,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,10240,0.033655110332700945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,16384,0.02624888883696662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,7168,0.01903555625014835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,512,0.009310222334331935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,65536,0.265209780799018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,2560,0.02143200072977278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,3072,0.006808888994985157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,1024,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,16384,0.020271110865804885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,6144,0.010915555887752108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,4096,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,4096,0.016859556237856548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,6144,0.03605688942803277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,32,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,3072,0.016206221448050607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,16384,0.04497777753406101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,3072,0.015043555034531487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,8192,0.016466667254765827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,7168,0.021689777572949726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,5120,0.02680088910791609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,256,0.0053048887186580235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,1024,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,65536,0.052172445588641696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,1024,0.009689778089523315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,1536,0.0274008893304401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,6144,0.015861334072219003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,6144,0.012368000215954252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,16384,0.022854222191704646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,4096,0.06548888815773858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,64,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,128,0.01479111115137736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,12288,0.04688622223006355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,8192,0.014867555763986377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,768,0.004541333350870344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,3584,0.02072622213098738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,8192,0.032320890161726214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,64,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,10240,0.05494933327039083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,128,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,65536,0.06598489152060615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,256,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,65536,0.22007289197709826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,7168,0.01991555591424306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,6144,0.013789333403110504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,1024,0.005969777703285217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,3072,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,6144,0.021336888273557026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,65536,3.756993611653646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,7168,0.020301333732075162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,3072,0.008909333083364699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,2560,0.007055111229419708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,128,0.009680888719028896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,7168,0.009659555223253038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,3584,0.01518311103185018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,10240,0.32069600952996147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,10240,0.36302844683329266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,1024,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,10240,0.12854133711920843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,768,0.005214222189452913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,32,0.008640888664457533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,768,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,7168,0.01032977799574534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,2048,0.007310222420427535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,768,0.028468443287743464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,64,0.006395555618736479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,8192,0.026124444272783067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,8192,0.02425688836309645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,65536,0.18569421768188477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,65536,0.40543378723992246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,32,0.0029884444342719186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,32,0.021742223037613764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,32,0.009368889033794403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,10240,0.35953688621520996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,4096,0.016673778494199116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,3584,0.1343066692352295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,2048,0.013833777772055732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,128,0.005510222166776657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,1536,0.006234666539563074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,7168,0.0076142218377855085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,32,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,65536,0.014098667436175875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,3072,0.05290844374232822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,8192,0.02973066767056783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,16384,0.019890666007995605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,3584,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,5120,0.010865777730941772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,6144,0.01657777859105004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,128,0.018209778600268893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,2560,0.006234666539563074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,768,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,512,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,512,0.006611555400821898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,256,0.004202666795916027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,10240,0.03264888789918687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,2048,0.014807111687130399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,2048,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,1536,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,7168,0.03015466531117757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,8192,0.14289955298105875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,16384,0.16581332683563232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,2560,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,64,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,768,0.003641777982314428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,10240,0.014671999547216626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,65536,0.4103199905819363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,2560,0.006350222147173352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,1024,0.00666133314371109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,10240,0.015432000160217285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,7168,0.18225510915120444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,64,0.007978666987684038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,7168,0.026359111070632935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,12288,0.017678222722477384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,65536,0.10952444208992852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,4096,0.008339555727110969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,1536,0.006312888943486744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,256,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,12288,0.025332444243960913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,1536,0.014150222142537435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,5120,0.012060444388124677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,8192,0.13335555129581028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,10240,0.016528000434239704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,32,0.00554044461912579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,128,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,16384,0.031574222776624895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,65536,0.009733333355850643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,1024,0.004566222015354368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,2048,0.028672890530692205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,32,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,3584,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,64,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,8192,0.011003555523024665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,6144,0.011605333122942181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,65536,0.07456888755162557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,8192,0.02747111188040839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,12288,0.018848000301255118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,64,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,768,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,2048,0.009651555783218807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,256,0.015457777513398064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,3072,0.01721333298418257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,1536,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,256,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,2560,0.01094044413831499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,3072,0.007548444800906711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,10240,0.03048355711830987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,16384,0.05015200045373705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,3072,0.01583377851380242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,2560,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,1536,0.008999110923873054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,6144,0.011408000356621213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,3072,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,1024,0.007500444849332173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,4096,0.02200711104604933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,128,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,2560,0.007687110867765214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,5120,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,4096,0.03641511003176371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,256,0.02344177828894721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,2560,0.022290666898091633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,8192,0.03153689040078057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,7168,0.026936888694763184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,65536,0.0758168896039327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,2048,0.005650666852792104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,4096,0.017639999588330586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,8192,0.0165928883685006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,16384,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,2048,0.009447111023796929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,5120,0.05555999941296048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,65536,0.08318844768736097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,16384,0.018971555762820773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,768,0.0041680000722408295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,10240,0.024031110935741003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,4096,0.012758221891191272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,65536,0.1877902216381497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,4096,0.03533777925703261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,256,0.01822933389080895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,768,0.005554666535721884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,256,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,3584,0.028579556279712256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,32,0.0029022221763928733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,4096,0.009112888740168678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,10240,0.017096888687875535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,128,0.014499555031458536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,2048,0.014135110709402295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,8192,0.019880000087949965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,768,0.003564444267087512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,8192,0.02201866606871287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,10240,0.013757333159446716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,5120,0.03052355514632331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,12288,0.025102222959200542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,12288,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,4096,0.01756977703836229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,3072,0.007307555940416124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,12288,0.1071573363410102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,8192,0.0229102224111557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,3584,0.014579556054539151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,1536,0.005982222242487802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,12288,0.041152000427246094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,7168,0.01202488856183158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,3584,0.006903999795516332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,768,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,2560,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,10240,0.02286400066481696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,768,0.00453333349691497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,2048,0.007671111159854465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,1536,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,6144,0.04587733414438036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,128,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,64,0.002789333255754577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,1536,0.017143999536832173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,10240,0.021568889419237774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,12288,0.04635111159748501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,1024,0.013758222262064615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,6144,0.01441333360142178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,4096,0.01366222235891554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,6144,0.028022223048739966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,1536,0.013175999952687157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,3072,0.02228533393806881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,3584,0.03193511234389411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,128,0.0031991110493739447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,16384,0.24212800131903756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,128,0.003921777837806278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,2560,0.00700444479783376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,65536,0.08134400182300143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,256,0.02025422288311852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,65536,0.11777067184448242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,1536,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,65536,0.12732533613840738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,1536,0.007655111451943715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,6144,0.015433778365453085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,3072,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,65536,0.13163288434346518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,3584,0.007748444047239091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,10240,0.033328887489106916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,1536,0.10375377866956924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,16384,0.029614223374260798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,8192,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,16384,0.021935110290845234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,2048,0.012119111087587146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,3584,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,10240,0.06419377856784396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,3072,0.06744266880883111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,65536,0.32782665888468426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,1024,0.013152889079517789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,10240,0.011734222372372946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,1024,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,16384,0.03820977939499749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,12288,0.03317155440648397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,4096,0.11447821723090278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,2048,0.03759110967318217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,1536,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,256,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,768,0.017440888616773818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,768,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,6144,0.007818666597207388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,12288,0.0292559994591607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,5120,0.15395732720692953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,2048,0.009185777770148383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,8192,0.027878221538331773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,64,0.011582222249772815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,128,0.01018666641579734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,10240,0.03710755705833435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,512,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,8192,0.0745146671930949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,768,0.017942221628295053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,12288,0.017316444052590262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,8192,0.0103502223889033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,65536,0.18460622098710802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,512,0.014856000741322836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,256,0.01015466699997584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,64,0.005205333232879639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,12288,0.03394044438997904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,2048,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,768,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,3072,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,128,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,6144,0.012181333369678922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,6144,0.050586667325761586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,6144,0.02299822204642826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,256,0.0052284445199701525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,16384,0.02458577851454417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,3072,0.00738666703303655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,64,0.08616444137361313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,768,0.01389600005414751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,65536,0.10651822222603692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,3072,0.015305777390797934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,128,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,2048,0.01789422167672051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,64,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,256,0.015106666419241162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,768,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,256,0.01514755520555708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,256,0.016920000314712524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,5120,0.0780151089032491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,256,0.024004444479942322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,1536,0.02199199961291419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,4096,0.00904622260067198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,256,0.005255110975768831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,3584,0.0631600022315979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,512,0.010449777874681683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,256,0.015154666370815702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,65536,0.877854241265191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,3072,0.014096889230940076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,6144,0.011115555961926779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,512,0.009850666754775578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,2560,0.013487999637921652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,16384,0.037921776374181114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,5120,0.028039998478359644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,65536,0.13015111287434897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,16384,0.02236088944805993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,1024,0.008336889247099558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,3072,0.05074488785531786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,8192,0.016925333274735343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,2560,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,1536,1.6244630813598633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,2560,0.01313333296113544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,3072,0.03629066546758016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,6144,0.015825778245925903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,32,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,512,0.06632977724075317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,4096,0.013420444395807056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,512,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,256,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,512,0.008960000342792934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,2560,0.01034044474363327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,4096,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,1024,0.010124444133705562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,128,0.019550222489568923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,768,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,32,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,3072,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,2560,0.011063999599880643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,2560,0.026544888814290363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,5120,0.013008889224794177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,128,0.019259555472267997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,128,0.005531555662552516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,768,0.017103999853134155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,3584,0.006847111301289664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,4096,0.024499555428822834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,2560,0.016487111647923786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,7168,0.04427999920315213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,12288,0.038104888465669416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,10240,0.01756444407833947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,1024,0.008299555215570662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,512,0.02014488809638553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,8192,0.01994222237004174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,128,0.05507111218240526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,6144,0.01648977730009291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,8192,0.012467555701732635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,65536,0.10435289144515991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,3584,0.01350311107105679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,10240,0.16040089395311144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,2048,0.006002666635645761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,5120,0.07789511150783963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,3072,0.0865315530035231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,128,0.0028453332682450614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,3072,0.009490666290124258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,512,0.003522666792074839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,12288,0.013111111190583972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,3584,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,8192,0.04599377844068739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,12288,0.12053955925835504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,3584,0.02124799953566657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,65536,0.3279208872053358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,2048,0.011429333024554782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,12288,0.049399998452928334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,2048,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,256,0.01516355574131012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,5120,0.01756177842617035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,10240,0.027726221415731642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,64,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,5120,0.027954667806625366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,1536,0.020357333951526217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,512,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,12288,0.040561778677834406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,10240,0.0654675563176473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,3584,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,16384,0.06102310948901706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,512,0.00488711107108328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,3072,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,128,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,512,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,7168,0.018217777212460835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,10240,0.018256889449225534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,2560,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,1536,0.010982222027248807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,5120,0.04742933313051859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,2048,0.005223999834722943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,65536,0.07655022541681926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,256,0.008291555775536431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,4096,0.010679110884666443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,1536,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,64,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,512,0.028961777687072754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,1024,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,1024,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,5120,0.03769955701298184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,2560,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,65536,0.08269333177142672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,4096,0.00833511104186376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,16384,0.08004266685909696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,1024,0.011746666497654386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,8192,0.010391111175219217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,16384,0.3129359881083171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,128,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,5120,0.017809778451919556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,512,0.004860444615284602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,4096,0.011039999624093374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,12288,0.03432088759210374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,16384,0.022327999273935955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,768,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,12288,0.016442666451136272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,64,0.0026951111439201566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,256,0.01677688956260681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,1536,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,16384,0.03184355629814996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,6144,0.009676444033781687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,256,0.00599644457300504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,12288,0.014668444792429606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,5120,0.011046222514576383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,512,0.014101333088344999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,3072,0.08479733599556817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,2048,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,128,0.01681511104106903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,65536,0.08466933170954387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,2048,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,768,0.038309332397249006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,768,0.009011555876996782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,2560,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,768,0.016898666818936665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,128,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,768,0.01854222185081906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,6144,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,5120,0.02349688940578037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,16384,0.03900711072815789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,512,0.013100444442696042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,16384,0.05588799715042114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,10240,0.012725333372751871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,3072,0.016653334101041157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,3072,0.15490844514634874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,5120,0.009342222577995723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,4096,0.012367111113336353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,10240,0.32770755555894643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,12288,0.0466542210843828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,5120,0.03907910982767741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,3584,0.01958933307064904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,6144,0.22949067751566568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,65536,0.06442755460739136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,7168,0.020280000236299302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,12288,0.03256622287962172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,10240,0.01511111193233066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,3584,0.008683555656009251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,1536,0.043954667117860585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,3584,0.011318222516112857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,7168,0.019545776976479422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,1024,0.014145778285132514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,7168,0.008279999925030602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,8192,0.01993688941001892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,128,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,5120,0.01926933394538032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,4096,0.017449777987268236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,8192,0.021623111433453027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,6144,0.026471111509535048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,3584,0.016235555211702984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,1024,0.014139556222491793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,5120,0.009004444711738164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,1024,0.007268444531493717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,6144,0.07155466741985746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,10240,0.03735733363363478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,8192,0.050527999798456825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,3072,0.011700444751315646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,16384,0.04151111178927951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,2048,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,1024,0.031161778502994116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,3072,0.035951998498704694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,2048,0.0162462227874332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,8192,0.010671999719407825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,16384,0.010682666467295753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,32,0.002797333316670524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,2048,0.012103999654452005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,32,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,2560,0.022924444741672937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,16384,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,3072,0.021584888299306233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,7168,0.010966222319338055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,4096,0.013563555147912769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,8192,0.012071111136012606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,512,0.022034666604465906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,768,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,2560,0.015448000695970325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,128,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,65536,0.15434577729966906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,6144,0.009956444303194681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,3072,0.01479466590616438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,2560,0.007333333293596904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,768,0.018932445181740653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,512,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,8192,0.011029333704047732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,1536,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,3072,0.02528800070285797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,16384,0.018937778141763475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,1536,0.00795733349190818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,16384,0.023081777824295893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,65536,0.1541164451175266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,1024,0.006183111005359226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,2048,0.032815112007988825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,65536,0.3265315426720513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,12288,0.06582399871614245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,1536,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,10240,0.02250755495495266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,6144,0.015816888875431485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,1024,0.006226666685607698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,16384,0.016184889607959323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,768,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,8192,0.07044800122578938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,512,0.02237155536810557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,7168,0.22940709855821398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,8192,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,3584,0.007309333317809635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,1024,0.007956444389290279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,6144,0.043256001340018384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,10240,0.021319111188252766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,12288,0.02921066681543986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,4096,0.10676711135440403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,10240,0.04532444477081299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,10240,0.04105066590838962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,2048,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,2560,0.08026755518383451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,16384,0.0325911111301846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,768,0.0035217776894569397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,1024,0.012433778080675336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,8192,0.011875555747085147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,6144,0.009628444082207149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,1536,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,1024,0.022075555390781824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,3072,0.012237333589129977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,16384,0.01588088936275906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,3584,0.010295111272070143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,65536,0.057256890667809375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,128,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,256,0.003504000190231535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,5120,0.02179199953873952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,1024,0.004879999905824661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,128,0.013470222552617392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,1024,0.006312888943486744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,128,0.002855111120475663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,768,0.007352888584136963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,32,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,10240,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,3584,0.007851555943489075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,1024,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,6144,3.059273825751411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,2560,0.010727110836240979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,32,0.00453599997692638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,1024,0.011348444554540845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,1536,0.007630222373538547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,2048,0.009749333063761393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,5120,0.01572355628013611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,10240,0.018999111321237352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,1536,0.014074666632546319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,256,0.015141333142916361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,1024,0.01276000009642707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,65536,0.055712891949547656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,10240,0.054245332876841225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,16384,0.036892443895339966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,512,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,1536,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,3584,0.010459555519951714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,16384,0.04148799843258328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,256,0.015207999282413058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,512,0.016533333394262526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,64,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,2560,0.009322666459613377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,1024,0.006462222172154321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,1536,0.004556444370084339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,3584,0.007679111427730984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,16384,0.03201599915822347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,6144,0.007347555624114142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,12288,0.03927288783921136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,512,0.0035502223504914176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,2048,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,7168,0.551146666208903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,12288,0.03176711002985636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,65536,0.014567111929257711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,3072,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,256,0.0041928887367248535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,4096,0.019319999549123976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,3072,0.05492977632416619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,1024,0.030737777551015217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,768,0.05454488926463657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,32,0.0028560000161329904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,3072,0.013716444373130798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,7168,0.017352888981501263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,64,0.004936888813972473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,10240,0.4401422341664632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,2048,0.07893511321809557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,1536,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,5120,0.02497244377930959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,256,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,32,0.006228444476922353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,1536,0.004544888933499654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,7168,0.020649777518378366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,768,0.015496888094478183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,2048,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,8192,0.0053048887186580235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,2560,0.10994666814804077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,3072,0.014840889308187695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,64,0.002883555574549569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,2048,0.08041689130995008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,6144,0.010680889089902243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,3072,0.015881778465376962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,8192,0.018750222192870248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,6144,0.09423644675148858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,10240,0.01070666644308302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,6144,0.03627644313706292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,768,0.029707554313871596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,12288,0.019333332777023315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,256,0.016511999898486666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,8192,0.01348355578051673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,4096,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,768,0.018124444617165458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,256,0.014855111638704935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,2560,0.015801777442296345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,512,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,512,0.014504889647165934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,8192,0.2005608876546224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,2048,0.010438222024175854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,1024,0.052128000391854175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,256,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,64,0.009648888475365108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,3584,0.011014222270912595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,768,0.016566221912701923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,3072,0.048546665244632296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,512,0.007037333316273159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,128,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,8192,0.020633776982625324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,10240,0.029418667157491047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,16384,0.02126577827665541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,12288,0.01827022267712487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,3072,0.01755644381046295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,32,0.004912888838185204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,128,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,65536,0.014919110470347934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,8192,0.011092444260915121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,6144,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,512,0.004524444540341695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,1536,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,12288,0.0366906656159295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,65536,0.27636978361341685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,256,0.015189333094490899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,5120,0.010072888599501716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,6144,0.03976622223854065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,768,0.12360000610351562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,2560,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,32,0.008377778033415476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,64,0.04070933328734504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,768,0.02069511181778378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,1024,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,6144,0.03298844562636481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,1024,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,5120,0.011414222419261932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,65536,0.10827200280295478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,128,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,1024,0.00833777752187517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,64,0.0028648889727062653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,16384,0.38227023018731016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,256,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,2048,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,1536,0.05497422152095371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,16384,0.02567911148071289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,3584,0.20436978340148926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,5120,0.008714666797055138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,2560,0.0057608890864584185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,6144,0.014540443817774454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,128,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,2048,0.012171555724408893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,1536,0.014089778065681458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,1536,0.06638489166895549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,128,0.010097777677906884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,512,0.013873777455753751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,2048,0.04262222184075249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,256,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,1536,0.04482044445143806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,5120,0.00868711123863856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,512,0.07661955886416964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,2560,0.013724444641007317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,1024,0.01205244412024816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,2048,0.1445608933766683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,256,0.003855111284388436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,8192,0.01793599956565433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,5120,0.02306844459639655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,4096,0.010630221830474006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,3072,0.01239999963177575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,2560,0.011374221907721626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,10240,0.030424889591005113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,8192,0.02093155516518487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,768,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,3072,0.013826666606797112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,1024,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,256,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,1024,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,2048,0.0236488895283805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,10240,0.0647031135029263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,1024,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,256,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,1536,0.02791822287771437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,3072,0.007436444362004598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,3072,0.02294844388961792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,256,0.011736888852384357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,16384,0.046817776229646474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,256,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,10240,0.26103199852837455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,256,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,8192,2.408128102620443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,5120,0.018578666779730055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,3072,0.007248889240953658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,16384,0.02167555524243249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,4096,0.007646222081449296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,5120,0.009748444788985783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,12288,0.033441778686311506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,512,0.024717332588301763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,5120,0.10292533371183608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,2560,0.01753777762254079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,1024,0.0042239998777707415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,128,0.018216888109842937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,16384,0.25169422891404897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,32,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,32,0.004222222086456087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,1536,0.01705511079894172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,512,0.05080088973045349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,3584,0.0200284438000785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,32,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,7168,0.03782577647103204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,7168,0.035440001222822405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,256,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,10240,0.01941244469748603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,768,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,8192,0.014095111025704278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,768,0.004527111020353106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,65536,0.08854933579762776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,64,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,2560,0.006568888823191325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,65536,0.09182311428917779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,3072,0.03129955463939243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,128,0.015117333994971381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,10240,0.016160888804329764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,10240,0.012103999654452005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,16384,0.1273848878012763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,5120,0.016027554869651794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,3072,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,4096,0.11489689350128174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,6144,0.020641777250501845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,32,0.006664000037643645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,1536,0.007826666865083907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,3072,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,128,0.011356444822417365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,1536,0.01239999963177575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,1536,0.00923733330435223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,512,0.06052533123228285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,65536,0.39388712247212726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,5120,0.10308000114228989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,16384,0.033751110235850014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,3072,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,2560,0.021596444977654353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,65536,0.160044444931878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,8192,0.012094222009181976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,6144,0.016504888733228047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,5120,0.018886221779717337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,2048,0.020568889048364427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,5120,0.015930665863884818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,8192,0.14408089054955378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,32,0.005883555445406172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,768,0.0038924444880750445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,3072,0.012285333540704517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,2560,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,6144,0.009319110876984065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,3072,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,7168,0.026597332623269822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,3072,0.020434666011068556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,6144,0.01648977730009291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,4096,0.025791111919615004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,3584,0.007603555917739868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,6144,0.02251200046804216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,3584,0.039157334301206805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,1024,0.007322666545708974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,65536,0.18944266107347277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,1536,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,10240,0.02883111106024848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,5120,0.008981333010726506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,4096,0.035218665997187294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,7168,0.010040000081062317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,2048,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,64,0.004190222256713443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,2048,0.014182221558358936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,3584,0.014817777607176038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,5120,0.03379733363787333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,10240,0.027150221996837195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,128,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,2048,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,65536,2.151744842529297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,10240,0.039706667264302574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,32,0.010058666268984476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,1536,0.1833991077211168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,768,0.0057297779454125305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,4096,0.028121779362360638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,16384,0.11057511303159927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,3584,0.007501333124107785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,65536,1.5903982586330836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,512,0.011239111423492432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,4096,0.027785778045654297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,8192,0.019655111763212416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,1024,0.004540444248252445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,128,0.006464889066086874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,2048,0.014145778285132514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,768,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,128,0.013589333328935834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,4096,0.15086132950252956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,12288,0.03933599922392104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,768,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,10240,0.02292799949645996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,1536,0.011196444431940714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,5120,0.13285956117841932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,32,0.006888888776302338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,4096,0.008637333081828224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,3584,0.014857777290874057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,6144,0.00905600024594201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,1536,0.012182222472296821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,4096,0.00794311116139094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,65536,0.10552710956997341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,3072,0.012181333369678922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,3584,0.008120888637171851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,8192,0.008285333712895712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,1024,0.013129777378506131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,16384,0.018671999375025432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,2560,0.006696888970004187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,768,0.01847644481394026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,7168,0.0631271137131585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,3072,0.02822222312291463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,768,0.01721333298418257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,2560,0.016184000505341422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,65536,0.08233066399892171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,2048,0.006787555499209299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,128,0.006498666687144174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,7168,0.030406223403082952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,4096,0.011046222514576383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,1024,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,65536,0.17182044188181558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,16384,0.022692445251676772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,2048,0.007916444705592262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,10240,0.03056710958480835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,1536,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,8192,0.011024889018800525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,2048,0.005689777847793367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,12288,0.023702222439977858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,128,0.014869333969222175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,512,0.025672889418072168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,1024,0.003933333274390963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,128,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,7168,0.012276444170210095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,1024,0.006184000107977126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,1536,0.03644533289803399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,6144,0.008955555657545725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,1024,0.00461777796347936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,1024,0.13224266635047063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,128,0.0028737777223189673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,4096,0.010273777776294285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,10240,0.031798223654429116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,1536,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,65536,0.5117057694329156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,1536,0.01514311134815216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,7168,0.01923733287387424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,5120,0.011578666667143503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,5120,0.07102577553855048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,768,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,3072,0.29376265737745494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,2048,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,128,0.014529777897728814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,12288,0.021723555194007024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,4096,0.07067644596099854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,1024,0.06651822063657972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,3072,0.01298577752378252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,16384,0.13170133696662054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,512,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,3584,0.00794044468137953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,768,0.00349866681628757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,256,0.022655111220147874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,65536,0.17506578233506945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,6144,0.00902844468752543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,1536,0.004582222137186262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,2048,0.06040177742640177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,12288,0.024362666739357844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,1024,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,12288,0.013111111190583972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,2560,0.009371555513805812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,1536,0.012111110819710625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,3584,0.011777777638700275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,3584,0.015815110670195687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,12288,0.017500445246696472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,768,0.01718933383623759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,16384,0.03155644403563605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,1536,0.01053600013256073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,512,0.005650666852792104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,3072,0.03678488731384277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,12288,0.2686613400777181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,32,0.006612444503439798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,32,0.0028026666906144884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,65536,0.09895288944244385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,1024,0.006562666760550604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,128,0.005554666535721884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,10240,0.027698665857315063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,3072,0.016543111867374845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,5120,0.0159599996275372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,7168,0.02085955606566535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,4096,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,2560,0.011377777490350934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,7168,0.013890667094124688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,1536,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,3584,0.04385866721471151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,8192,0.03267466690805223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,2560,0.013507555756303998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,3072,0.012998222476906247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,128,0.011416888899273343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,6144,0.009202666580677032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,10240,0.04320089022318522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,4096,0.03744355506367154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,64,0.0028560000161329904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,128,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,10240,0.5797991222805446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,3584,0.010043555663691627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,8192,0.13297155168321398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,768,0.006531555619504716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,8192,0.023765333824687537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,768,0.047516445318857826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,3072,0.027462222509913977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,1536,0.006065777606434292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,32,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,6144,0.009552888572216034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,32,0.008947555389669206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,3584,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,1024,0.003866666720973121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,65536,0.15251733197106257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,7168,0.012102222277058495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,8192,0.016171556380059984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,1536,0.00654666663871871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,2560,0.023315555519527856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,1024,0.012727110750145383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,10240,0.01771022213829888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,4096,0.030964444080988567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,128,0.016720000240537856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,7168,0.33460177315606016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,12288,0.06706844435797797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,12288,0.0159288893143336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,3584,0.018216888109842937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,5120,0.14421244462331137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,256,0.015072888798183866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,1536,0.007368889119890001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,3584,0.007804444266690149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,1024,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,32,0.0028595555987623003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,3072,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,1024,0.009315555294354757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,65536,0.20527377393510607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,768,0.0041288890772395665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,3584,0.023131555981106226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,6144,0.00925688859489229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,10240,0.01738133364253574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,7168,0.019903111788961623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,7168,0.010004444254769219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,1536,0.014722666806644864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,4096,0.023039110832744177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,12288,0.03367822368939718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,1024,0.007278222176763747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,4096,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,512,0.021219554874632094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,3072,0.018872888551818002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,8192,0.1691208945380317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,7168,0.020284444093704224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,4096,0.01924888955222236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,4096,0.013599110974205865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,64,0.00564533347884814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,3584,0.013800000150998434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,768,0.009541333549552495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,32,0.003003555453485913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,3584,0.018651554981867473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,128,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,7168,0.04985510971811083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,7168,0.01960000064637926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,3584,0.007646222081449296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,12288,0.02166844407717387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,10240,0.009002666506502364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,256,0.005552000055710475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,4096,0.05034133460786608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,6144,0.01497866710027059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,7168,0.012064889073371887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,4096,0.00794311116139094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,5120,0.027440889014138117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,10240,0.02289955483542548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,2048,0.04909600151909722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,10240,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,4096,0.010309333602587381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,6144,0.009043555292818282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,128,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,128,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,65536,1.3162337409125435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,5120,0.10751377873950535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,5120,0.09413422478569879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,256,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,12288,0.013076444466908773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,3584,0.006135111053784688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,65536,1.0427138010660808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,256,0.012054222325483957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,12288,0.028791109720865887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,3584,0.00702311098575592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,7168,0.009986666341622671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,5120,0.01792533364560869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,3072,0.011756444142924415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,4096,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,256,0.015453333655993143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,7168,0.013321778012646569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,10240,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,3072,0.015556444724400839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,512,0.0127688886390792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,3584,0.01685066686736213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,6144,0.03307288885116577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,16384,0.019181332654423185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,65536,0.01481066644191742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,65536,0.4130942291683621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,4096,0.017775999175177682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,6144,0.018758222460746765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,512,0.009152889251708984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,6144,0.01404622197151184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,256,0.0051893335249688895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,6144,0.02812088860405816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,7168,0.02132622235351139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,6144,0.019349333312776353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,1536,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,1536,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,4096,0.013437333206335703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,512,0.0038897775941424896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,3584,0.0096133334769143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,5120,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,10240,0.055778668986426465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,64,0.009335111412737105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,5120,0.017328888177871704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,64,0.006671111202902264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,128,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,6144,0.007343110938866933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,64,0.0029271110478374693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,256,0.006856888946559694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,1024,0.005764444255166584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,1024,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,3584,0.010418666733635796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,1536,0.0090524446633127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,2048,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,1536,0.005009777843952179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,1536,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,5120,0.008752000000741746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,128,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,512,0.013776889277829064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,1536,0.018240888913472492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,512,0.02871555421087477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,4096,0.007919111185603673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,16384,0.01925511161486308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,5120,0.011749332977665795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,3584,0.08491733339097764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,3072,0.007387555307812161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,256,0.01688799924320645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,64,0.0031395554542541504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,1536,0.01276177747382058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,64,0.007920888562997183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,2560,0.007652444144090016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,1536,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,7168,0.013434666726324292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,2560,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,3584,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,64,0.014861333701345654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,2048,0.012168889244397482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,256,0.005223111146026188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,7168,0.016208888755904306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,3584,0.0123831108212471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,16384,0.03333866596221924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,1536,0.06025422281689114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,2048,0.006817777951558431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,3584,0.008667555948098501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,6144,0.11195200019412571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,7168,0.009377777576446533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,6144,0.02551377813021342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,8192,0.011364444262451597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,1536,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,2560,0.022620444496472675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,64,0.006923555499977536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,1536,0.006572444405820634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,10240,0.8311057620578342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,1024,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,256,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,3584,0.022318222456508215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,65536,0.13571200105879042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,12288,0.021487111846605938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,6144,0.013457777599493662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,7168,0.4293644428253174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,1536,0.018591110905011494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,1536,0.008976889153321585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,2048,0.0188737776544359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,65536,0.09434577491548325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,32,0.003014222201373842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,5120,0.02446311049991184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,2048,0.006836444553401735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,2560,0.015839111473825242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,5120,0.030565334690941706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,5120,0.018910222583346896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,5120,0.022616000639067754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,3584,0.021194666624069214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,16384,0.012079999678664737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,8192,0.024711999628278945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,2048,0.006725333217117522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,64,0.003488000068399641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,10240,0.03501866592301263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,5120,0.08942488829294841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,4096,0.018945778409639995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,6144,0.015095111396577625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,64,0.010296888649463654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,7168,0.02807733416557312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,65536,0.15174578295813665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,64,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,128,0.01550222271018558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,2560,0.04891555507977804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,512,0.01718311177359687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,8192,0.3058319886525472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,5120,0.010311110979980892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,65536,0.10455999771753947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,2560,0.028679112593332928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,5120,0.009332444104883406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,64,0.003264888914095031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,256,0.0031644445326593188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,1024,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,3072,0.013051555388503604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,3584,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,1536,0.030983110268910725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,12288,0.020764445265134174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,768,0.008338666624493068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,768,0.006912888752089606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,768,0.05219555563396878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,1024,0.0038897775941424896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,6144,0.01552088889810774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,4096,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,2560,0.1600115564134386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,4096,0.023078221413824294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,3072,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,5120,0.008356444537639618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,5120,0.024918221765094336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,768,0.013783999615245394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,256,0.00553955551650789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,4096,0.019268444842762418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,5120,0.018637332651350234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,2048,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,256,0.0031946665710873077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,3072,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,6144,0.018882667024930317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,5120,0.01648177703221639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,4096,0.01648177703221639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,2560,0.008263111114501953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,128,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,256,0.028431110911899145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,1024,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,16384,0.028639998700883653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,65536,0.06522222359975179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,128,0.00565422202150027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,3072,0.02216000027126736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,65536,0.28969422976175946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,16384,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,65536,0.14812089337242976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,1536,0.019746666153271992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,16384,0.024493333366182115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,4096,0.014869333969222175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,512,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,1536,0.022853333089086745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,256,0.016179554992251925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,8192,0.048804445399178394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,3072,0.009639999932712978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,32,0.007596444752481248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,8192,0.05788889196183947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,8192,0.013079110946920184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,7168,0.009690666364298927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,256,0.006433777925040986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,8192,0.02305511136849721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,768,0.020614221692085266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,2048,0.03165333469708761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,8192,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,1536,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,256,0.010344888601038191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,10240,7.228889465332031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,1024,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,1024,0.007624000310897827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,512,0.0052391112678580815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,5120,0.021033777130974665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,4096,0.012040889097584618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,8192,0.04051288962364197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,512,0.017833777599864535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,65536,0.08081066608428955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,16384,0.07611378033955891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,3072,0.013458666702111563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,3072,0.0069511110583941145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,16384,0.02038755516211192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,16384,0.020469332734743755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,16384,0.06486933098898993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,64,0.003289777785539627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,8192,0.014595554934607612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,64,0.006996444529957241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,6144,0.01902755598227183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,2048,0.007893333004580604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,3584,0.0359919998380873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,4096,0.26131733258565265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,3584,0.011222222612963783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,10240,0.014865777558750577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,3584,0.00982844415638182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,2560,0.019298666053348117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,256,0.0032106666929192017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,64,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,768,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,1536,0.019362666540675692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,1024,0.004102222207519743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,8192,0.06193333201938205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,1024,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,2048,0.011305777562989129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,4096,0.007583111524581909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,4096,0.012472888661755456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,65536,0.0654239985677931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,5120,0.1707271072599623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,3584,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,2560,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,512,0.0035537779331207275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,5120,0.012396444049146442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,2560,0.009082666701740688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,1536,0.07632088661193848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,12288,0.014892444014549255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,128,0.014633778068754407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,2560,0.012858666479587555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,7168,0.04493777619467842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,64,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,7168,0.009770666559537252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,128,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,128,0.004939555707905027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,8192,0.04287999868392944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,16384,0.031202667289310034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,12288,0.1599280039469401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,1536,0.04767022199100918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,6144,0.06770133309894137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,3584,0.007308444215191736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,10240,0.0259253333012263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,2560,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,1024,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,768,0.017666666044129264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,1536,0.014894222219785055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,12288,0.017895999881956313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,64,0.007358222372002072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,5120,0.008927110996511247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,32,0.03422844409942627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,3072,0.013884444203641681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,768,0.029378665818108454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,65536,2.044017791748047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,128,0.013153778182135688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,1536,0.012505778008037143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,2048,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,4096,0.04155200057559543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,65536,0.12633332941267225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,256,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,1024,0.00908088849650489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,768,0.0035822221802340615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,8192,0.08492177724838257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,1536,0.004583999928500918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,768,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,128,0.0148062225845125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,8192,0.019648889700571697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,2048,0.006895111252864202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,1024,0.014565333724021912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,512,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,1024,0.0038844446341196695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,2560,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,512,0.014661333627170987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,768,0.019164444671736825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,3584,0.07273688581254747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,128,0.017599110802014668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,1024,0.011422221859296164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,4096,0.011752888560295105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,12288,0.04386933313475715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,64,0.011711999773979187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,8192,0.012520000338554382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,1536,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,32,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,16384,0.04083733426200019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,65536,0.04561688833766513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,1536,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,4096,0.02470933397610982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,1536,0.011697777443461947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,256,0.0034711110509104202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,65536,0.08483378092447917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,3072,0.012532444463835822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,1536,0.008629333641793992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,2048,0.1222204499774509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,512,0.013517333401574029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,65536,0.13083555963304308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,6144,0.014446222119861178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,128,0.005256888767083486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,8192,0.029302222861184016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,6144,0.04223555657598707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,128,0.021589333812395733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,3584,0.02327555583582984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,1024,0.005982222242487802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,32,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,16384,0.14754399988386366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,2048,0.029397332006030615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,1024,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,6144,0.06447733110851712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,2048,0.012744888663291931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,512,0.003532444437344869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,512,0.0042444442709287005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,5120,0.016200888488027785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,10240,0.026165333059098985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,65536,0.06270400020811293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,8192,0.025306666890780132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,10240,0.02380977736579047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,32,0.002893333426780171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,16384,0.03623466690381368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,1024,0.004219555606444676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,2560,0.05798400110668606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,4096,0.1485795577367147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,128,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,10240,0.0191600008143319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,128,0.0052328887912962176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,2048,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,1024,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,1024,0.011728888584507836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,32,0.0158897770775689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,1024,0.046894222497940063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,256,0.009744889206356471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,5120,0.013793778088357715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,256,0.005478222337034013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,5120,0.029892444610595703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,8192,0.01997244523631202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,65536,0.05420444409052531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,2048,0.008440000315507254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,8192,0.008364444805516137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,32,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,512,0.005528888768619961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,3584,0.00998399986161126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,2560,0.015198222464985318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,512,0.009242666264375051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,2560,0.05520266625616285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,2560,0.009015999734401703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,7168,0.021165332860416834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,4096,0.016428444120619033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,10240,0.016032889485359192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,8192,0.015136000182893542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,256,0.005703111075692707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,1536,0.06616089079115126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,1024,0.019278221660190158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,2560,0.02088977727625105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,8192,0.024753777517212763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,2048,0.024756444825066462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,65536,0.1441511048210992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,4096,0.019828443725903828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,2560,0.04288622074657016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,12288,0.012572444147533841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,256,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,512,0.01442488862408532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,2048,0.00563733321097162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,768,0.01902844508488973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,7168,0.009211555951171452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,16384,0.014673777752452426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,128,0.005225777626037598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,5120,0.010036444498433007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,12288,0.02272622287273407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,128,0.04279199904865689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,4096,0.044658667511410184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,2560,0.01755822201569875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,2560,0.07365600268046062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,2560,0.007034666836261749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,3584,0.006992888947327931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,1536,0.009016888837019602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,128,0.003536888708670934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,2560,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,10240,0.01478488908873664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,3072,0.021988444858127173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,5120,0.011353777514563667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,6144,0.013183111117945777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,256,0.12051644590165879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,6144,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,8192,0.013802666631009845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,1024,0.0058755555914507965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,7168,0.01276266657643848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,3584,0.032243556446499295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,256,0.02201066745652093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,2048,0.009750222166379293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,3072,0.009397333694828881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,3072,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,512,0.0038453332251972626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,4096,0.008792888787057664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,2560,0.02851822310023838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,7168,0.5582248899671767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,16384,0.2799146705203586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,4096,0.027775999572541978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,2048,0.017492444978819955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,512,0.004680888934267892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,2560,0.008078222473462423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,2048,0.006691555596060223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,12288,0.3548151122199164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,1024,0.0120666664507654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,10240,0.04331377810902066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,12288,0.13030400541093615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,32,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,12288,0.013392888837390475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,512,0.006698666761318843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,7168,0.06067466735839844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,10240,0.011784888803958893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,128,0.0032115555885765287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,2560,0.012058667010731168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,2560,0.013730666703648038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,1024,0.013463111387358772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,1536,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,3584,0.009018667042255402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,256,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,2560,0.02407288882467482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,512,0.006344888773229387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,1536,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,3584,0.023728888895776536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,3072,0.009378666679064432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,12288,0.08346933126449585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,768,0.014545778433481852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,4096,0.022100443641344707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,256,0.012871111432711283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,512,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,256,0.014103111293580798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,16384,0.03048266636000739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,64,0.006774222271309958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,1024,0.00599911105301645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,12288,0.0863422221607632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,256,0.1294177770614624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,4096,0.05176977647675408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,1536,0.005214222189452913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,12288,0.04009066687689887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,12288,3.647533416748047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,8192,0.02366399930583106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,65536,3.2596087985568576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,7168,0.213499559296502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,8192,0.036511109934912786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,6144,0.34926756223042804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,10240,0.027561777167850073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,16384,0.016150222884284127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,6144,0.00868888861603207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,128,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,6144,0.014821334017647637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,2048,0.013373333546850415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,768,0.012800888882742988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,1024,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,2560,0.029847999413808186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,12288,0.10280711121029323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,64,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,16384,0.04691466689109802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,3072,0.007637333538797166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,1536,0.01091288857989841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,3072,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,16384,0.024318221542570326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,2048,0.01757777730623881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,16384,0.018532445033391316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,10240,0.016876444220542908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,1536,0.008236444658703273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,1024,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,65536,0.09391644265916611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,65536,0.062252442042032875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,10240,0.015957333975368075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,65536,0.5233395364549425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,512,0.03035022152794732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,3072,0.024691555235120986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,256,0.014084445105658637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,4096,0.00830577810605367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,512,0.033886220720079206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,10240,0.00903911143541336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,6144,0.01185066666867998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,64,0.0028106667515304354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,2048,0.011925333076053195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,65536,0.16757333278656006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,10240,0.07222311364279853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,12288,0.10231733322143555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,3072,0.01588444411754608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,1536,0.00958933350112703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,12288,0.011528888510333167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,256,0.012421333127551608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,768,0.003855111284388436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,64,0.01754044493039449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,65536,0.06528977553049724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,6144,0.024512888656722173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,10240,0.03352800011634827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,3584,0.008272888759771982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,512,0.013770666387346057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,1536,0.01165688865714603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,4096,0.026148445076412622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,64,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,2560,0.008033778104517195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,5120,0.01638488968213399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,6144,0.015055111712879606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,4096,0.01096088853147295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,2560,0.022704000274340313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,65536,0.05547733439339531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,10240,0.016143111719025504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,1024,0.004219555606444676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,8192,0.010789333118332757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,4096,0.010649777948856354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,3072,0.00832088871134652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,12288,0.023585778143670823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,10240,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,128,0.004879111217127906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,32,0.003871110992299186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,4096,0.00833777752187517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,512,0.006802666518423293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,4096,0.011554666691356234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,65536,0.19622933864593506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,4096,0.01829688913292355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,65536,0.702946662902832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,8192,0.01754666699303521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,65536,0.07530488570531209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,3072,0.0074551113777690465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,128,0.0035422220826148987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,8192,0.012496888637542725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,512,0.004146666576464971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,16384,0.021570665968788996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,3072,0.007334222396214803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,4096,0.007849777738253275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,256,0.0064222220745351576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,2560,0.01622577839427524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,128,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,5120,0.016805332567956712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,16384,0.030223998758527968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,5120,0.024572443630960252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,8192,0.039876444472206965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,128,0.0031759999692440033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,3584,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,4096,0.008417777717113495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,3584,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,5120,0.01721244388156467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,5120,0.14974133173624674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,64,0.03150755498144362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,8192,0.012431110772821637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,1024,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,256,0.011713777979214987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,16384,0.5554800033569336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,32,0.004915555732117759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,3072,0.01954844428433312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,768,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,1536,0.013759999639458127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,5120,0.4600808885362413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,2560,0.02014044423898061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,5120,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,16384,0.02092444399992625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,1536,0.01833511061138577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,512,0.0034648889882696998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,256,0.003544888976547453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,7168,0.02090488870938619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,1024,0.009364444348547194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,12288,0.03183999988767836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,256,0.0031928889867332247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,6144,0.05795200003517998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,6144,0.04922577738761902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,10240,0.016933333542611863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,2048,0.014162666267818876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,512,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,1024,0.014093332820468478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,4096,0.010048000348938836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,2048,0.015171556009186639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,16384,0.0429368880059984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,1536,0.016007110476493835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,2048,0.008797333472304875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,8192,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,16384,0.024319110645188227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,1024,0.06779111093944974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,256,0.005209777918126848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,1536,0.007789333661397298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,5120,0.018170666363504197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,65536,0.05472088853518168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,1024,0.006773333168692059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,128,0.014872888724009195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,6144,0.019024888674418133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,4096,0.08535999721950954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,64,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,5120,0.01818933255142636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,12288,0.019461333751678467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,32,0.004344888859324985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,3072,0.011362666885058085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,5120,0.013129777378506131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,5120,0.008018666671382057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,4096,0.00923288861910502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,32,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,3584,0.07977688974804349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,5120,0.026005332668622334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,256,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,16384,0.02197244432237413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,12288,0.035387555758158364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,2560,0.008624888956546783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,16384,0.35002843538920086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,8192,0.014708444476127625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,1024,0.006888888776302338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,12288,0.8164826499091254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,3072,0.010849778023031024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,512,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,256,0.016873778568373788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,64,0.007881777981917063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,256,0.0053031109273433685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,4096,0.008864888714419471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,1024,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,3584,0.011493333511882357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,4096,0.018585777944988675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,64,0.002862222285734283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,768,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,2048,0.011057777537239922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,8192,0.02168088820245531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,12288,0.014174222946166992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,768,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,256,0.0032720000793536506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,2048,0.013930666777822705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,8192,0.12089422014024521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,32,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,1536,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,512,0.0069271110826068455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,2048,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,1024,0.005378666851255629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,256,0.0061039999127388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,3072,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,10240,0.027980443504121568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,768,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,12288,0.02281155520015293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,128,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,3072,0.057743999693128795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,16384,0.03057066599527995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,768,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,12288,0.016840888394249808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,768,0.017528888252046373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,2560,0.010988444089889526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,3072,0.008791999684439765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,256,0.006520888871616787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,2560,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,1536,0.008355555435021719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,1536,0.010715555813577441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,3584,0.01891911029815674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,1536,0.009204444785912832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,1536,0.013658666776286231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,6144,0.012120889292822944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,128,0.01110666659143236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,768,0.0038577777643998465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,128,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,12288,0.010797333386209277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,65536,0.08532266484366523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,7168,0.18596000141567656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,10240,0.02500711050298479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,3072,0.008614222208658854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,12288,0.014096889230940076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,16384,0.03820800118976169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,32,0.009639999932712978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,65536,0.05681600173314413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,5120,0.013799111048380533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,32,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,8192,0.1559271150165134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,3584,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,12288,0.020809777908855014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,1536,0.01329066687160068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,5120,0.025447110335032146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,1536,0.008277333445019191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,6144,0.2597111066182454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,16384,0.0378479990694258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,32,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,6144,0.023528888821601868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,1536,0.004575110971927643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,2048,0.01925511161486308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,1024,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,12288,0.0900613334443834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,6144,0.15008355511559382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,2560,0.006990222467316522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,7168,0.030568000343110826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,128,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,3072,0.017847110827763874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,7168,0.012305777933862476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,2560,0.012799999780125089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,32,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,128,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,3072,0.006525333142942852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,2560,0.03698044353061252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,3584,0.012433778080675336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,64,0.0027226667023367355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,1024,0.023000000251664057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,7168,0.011172444456153445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,10240,0.011920889218648275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,3072,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,8192,0.015116444892353482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,10240,0.011411555939250521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,3584,0.021533333592944678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,4096,0.015807999504937064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,4096,0.06958488623301189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,2048,0.018534221582942538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,6144,0.011064888702498542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,64,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,8192,0.01593244406912062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,768,0.007269333634111617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,1536,0.008799110849698385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,2560,0.016424889365832012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,2560,0.04205866654713949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,3584,0.009498666558000777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,256,0.01516888870133294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,6144,0.035051554441452026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,16384,0.03408266769515143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,3584,0.0421048899491628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,3584,0.0170133329100079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,8192,0.01037599974208408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,1024,0.0037679999238914917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,8192,0.032496889432271324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,2048,0.01311377767059538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,64,0.003567999849716822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,65536,0.017893332574102614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,65536,0.14264444510142008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,6144,0.02724444369475047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,64,0.008788444101810455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,16384,0.23010577095879448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,3584,0.019182221757041085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,256,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,512,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,64,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,1024,0.053620444403754346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,128,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,32,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,512,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,32,0.01223288890388277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,512,0.011406222151385413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,65536,5.981394873725043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,128,0.013825777504179211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,10240,0.027095999982621934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,10240,0.011047110789351992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,3584,0.11331378089057074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,3584,0.06580355432298449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,2560,0.07797600163353814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,65536,1.0710532930162218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,2048,0.019145778483814664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,128,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,5120,0.019333332777023315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,256,0.005225777626037598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,2048,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,1536,0.015448000695970325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,6144,0.009308444129096137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,3072,0.020911999874644812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,1536,0.008662222160233391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,3072,0.015104888214005364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,128,0.015863110621770222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,3072,0.010967999696731567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,12288,0.016196444630622864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,1536,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,128,0.005469333380460739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,6144,0.009038222332795462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,32,0.00795911086930169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,12288,0.01685244507259793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,3584,0.015241778559154935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,2560,0.006960888703664144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,7168,0.02622933354642656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,6144,0.027115555273161993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,8192,0.008943110704421997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,12288,0.01993777851263682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,32,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,128,0.0032177778581778207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,8192,0.024930665890375774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,8192,0.010677333507272931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,128,0.01479555500878228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,16384,0.04730755421850416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,8192,0.08864088853200276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,6144,0.009686222506894005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,1024,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,7168,0.013462222284740873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,256,0.0037173334923055437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,10240,0.012768000364303589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,3072,0.010695999695195092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,768,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,16384,0.08432177702585857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,4096,0.05243733194139269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,12288,0.0317795541551378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,768,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,256,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,2048,0.016184889607959323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,8192,0.02405333353413476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,5120,0.013106666505336761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,12288,0.041497776905695595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,1536,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,2560,0.009476444787449306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,128,0.014846222268210517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,256,0.018924444913864136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,7168,0.11816267172495525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,3072,0.03628977802064683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,2560,0.008348444269763099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,16384,0.020999110407299463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,8192,0.02421599957678053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,128,0.005254222287072076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,7168,0.03130133284462823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,16384,0.07716800106896295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,6144,0.012350222302807702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,1024,0.08247822523117065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,8192,0.01179377817445331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,4096,0.08411644564734565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,16384,0.02236888971593645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,3584,0.032285332679748535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,256,0.01682044400109185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,12288,0.057018664148118764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,768,0.014813333749771118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,7168,0.02569599946339925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,8192,0.02968177861637539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,512,0.007686222592989604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,16384,0.036399109496010676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,10240,0.012746666868527731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,256,0.0166275550921758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,4096,0.01717422240310245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,512,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,10240,0.025798221429189045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,128,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,512,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,512,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,768,0.0041857775714662345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,4096,0.01645155582163069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,1024,0.007618666523032718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,3584,0.020230222079488967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,2560,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,256,0.019864888654814828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,768,0.003900444342030419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,1536,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,3072,0.008723555339707268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,3584,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,8192,0.012847999731699625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,1536,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,128,0.00518488883972168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,8192,0.06295822064081828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,65536,0.06243822309705946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,16384,0.04295377930005392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,2560,0.013814222481515674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,6144,0.034973333279291786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,7168,0.029290666182835896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,1536,0.014780445231331719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,768,0.016881777180565726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,2560,0.008634666601816813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,3072,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,10240,0.01753777762254079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,2560,0.045125332143571645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,10240,0.23689511087205675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,6144,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,2560,0.008753777378135258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,4096,0.21992622481452095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,2560,0.011190222369299995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,7168,0.014737778239780001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,3584,0.015207999282413058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,2048,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,768,0.005764444255166584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,7168,0.10008622540367974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,128,0.014485332700941296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,10240,0.035841776265038386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,3584,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,256,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,3584,0.01718133356836107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,7168,0.1574595504336887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,12288,0.031161778502994116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,3072,0.007261333366235097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,1024,0.023817777633666992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,6144,0.011808888779746162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,1024,0.012446222205956778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,3584,0.06446133057276408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,10240,0.02160800000031789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,8192,0.02230666743384467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,128,0.00793422261873881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,7168,0.24323111110263398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,512,0.018380444910791185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,768,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,512,0.0032222221295038858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,8192,0.009339555270142024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,512,0.006609777609507243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,2560,0.0952133337656657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,3584,0.042303111818101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,16384,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,256,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,12288,0.04002133342954848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,2048,0.07216444280412462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,512,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,768,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,256,0.015111999379263984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,512,0.012783110969596438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,4096,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,65536,0.26451288329230416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,1024,0.09801066584057277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,3072,0.013470222552617392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,2560,0.012163555456532372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,1024,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,2560,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,2048,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,65536,0.1900097794002957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,512,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,7168,0.038388444317711726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,16384,0.36193511221143937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,16384,1.5595760345458984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,768,0.016889777448442247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,5120,0.04605244596799215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,6144,0.027658666173617046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,768,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,768,0.12261866198645698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,12288,0.10228355725606282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,3584,0.026943110757403906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,2560,0.009450666606426239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,256,0.014779556128713818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,1024,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,768,0.020404444800482858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,3072,0.01478666729397244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,16384,0.13286933634016249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,32,0.005680888891220093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,16384,0.05134488807784187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,6144,0.008926221893893348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,2560,0.015032889114485847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,16384,0.0210479994614919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,7168,0.010040000081062317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,64,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,5120,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,16384,0.4654960102505154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,7168,0.1115422248840332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,8192,0.09793777598275079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,128,0.010337777435779572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,32,0.016883555385801528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,64,0.01685333251953125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,7168,0.026072000463803608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,128,0.030255999830034044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,65536,0.0634711119863722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,12288,0.0236817780468199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,4096,0.037900444534089826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,768,0.004858666823969947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,1024,0.00867555538813273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,4096,0.034860445393456355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,12288,0.07787111070421007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,256,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,8192,0.018615111708641052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,3072,0.008685333033402761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,128,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,3584,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,2560,0.009711999860074785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,5120,0.03297955460018582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,65536,0.10159199767642552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,128,0.014481777946154276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,8192,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,512,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,3584,0.035440001222822405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,32,0.08799199925528632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,12288,0.062485330634646945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,1536,0.02288888891537984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,10240,0.03014666504330105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,16384,0.5044817924499512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,3584,0.00979288915793101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,2560,0.024145777026812237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,65536,0.10017689069112141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,6144,0.009370666411187913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,512,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,2048,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,6144,0.04551466637187534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,65536,0.0932791100607978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,6144,0.12662222650316027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,6144,0.023591111103693645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,6144,0.011726222104496427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,5120,0.020247111717859905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,2560,0.0417137775156233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,128,0.00655377780397733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,6144,0.02466222147146861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,16384,0.008557333714432186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,6144,0.009593778186374241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,4096,0.016178665889634024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,512,0.01882755590809716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,3584,0.013009777499569787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,1024,0.0339662233988444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,6144,0.009001777403884465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,32,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,10240,0.015277332729763456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,16384,0.02747822304566701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,64,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,5120,0.00794488853878445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,10240,0.014739554789331226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,8192,0.023407111565272015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,3072,0.015485333071814643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,256,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,128,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,4096,0.008544888761308458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,3584,0.021088000800874498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,3584,0.010681777364677854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,5120,0.2863608996073405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,1536,0.09824177953932021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,3072,0.02365422248840332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,6144,0.013111111190583972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,1024,0.004230222354332606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,12288,0.056932442718082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,128,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,1024,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,2048,0.011126221881972419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,16384,0.212900439898173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,65536,0.07264444563123915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,7168,0.012070222033394707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,2048,0.030528889762030706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,32,0.006573333508438534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,16384,0.03291733397377862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,10240,0.011758222348160215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,1536,0.05056711037953695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,5120,0.02607911162906223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,8192,0.3639644516838922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,64,0.004940444396601783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,10240,0.020290666156344943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,10240,0.03304355674319797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,2048,0.013462222284740873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,8192,0.14504622088538274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,1024,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,16384,0.022462222311231825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,256,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,7168,0.014208889669842191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,8192,0.011409777734014722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,16384,0.04459911253717211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,12288,0.014183110660976835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,3584,0.031967110104031034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,12288,0.1117395559946696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,2048,0.015849777393870883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,3584,0.016332444217469957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,2048,0.016380444169044495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,5120,0.033394667837354876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,2560,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,256,0.003165333428316646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,4096,0.01616533266173469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,32,0.0031582222630580268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,16384,0.021545777718226116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,768,0.01683288812637329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,7168,0.020288000504175823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,2048,0.01684977776474423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,2560,0.008648888932334052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,5120,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,5120,0.010668444136778513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,768,0.028986665937635634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,128,0.014502222339312235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,6144,0.021993777818149988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,1024,0.0057591112951437635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,4096,0.024616888827747766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,7168,0.028202666176689997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,3584,0.09009866582022773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,32,0.05452355411317614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,2560,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,7168,0.008967111508051554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,128,0.014832889040311178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,10240,0.010264000131024255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,1024,0.0052177777720822235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,16384,0.031493332650926374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,1536,0.02514222264289856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,65536,0.24660889307657877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,6144,0.01862844493654039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,10240,0.02772533396879832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,6144,0.013224889006879596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,8192,0.010760889285140567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,128,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,1536,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,3584,0.059097780121697314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,256,0.006548444430033366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,1024,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,1536,0.028498666154013738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,768,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,768,0.01754577789041731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,7168,0.013936888840463428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,6144,0.020470221837361652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,16384,0.016167110866970487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,2560,0.011022222538789114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,256,0.015134221977657743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,12288,0.017668444249365065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,3584,0.009695111049546136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,5120,0.008649778034951951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,65536,0.22455644607543945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,12288,0.06447466876771715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,128,0.0063075555695427795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,128,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,7168,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,128,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,5120,0.023623110519515142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,10240,0.028699556986490887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,2048,0.01570222278436025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,768,0.016901334126790363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,256,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,256,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,16384,0.02991644541422526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,3072,0.03179733289612664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,3072,0.01622488929165734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,4096,0.02218222286966112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,12288,0.38517067167494035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,1536,0.08992266654968262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,512,0.009358222285906473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,3584,0.11932355827755398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,512,0.009979555176364051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,512,0.03475644522243076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,512,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,8192,0.02567200031545427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,2048,0.026782222919993933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,128,0.007896000312434303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,3072,0.19047021865844727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,768,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,2048,0.013042666845851474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,3072,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,128,0.002793777734041214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,512,0.007333333293596904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,1024,0.014514666464593677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,128,0.021635555558734473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,2560,0.006480000085300869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,3584,0.010616000327799054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,6144,0.029050668080647785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,6144,0.031253334548738264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,16384,0.022014222211307947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,16384,0.055480887492497764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,8192,0.04039288891686334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,128,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,5120,0.040743112564086914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,1024,0.0076142218377855085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,3584,0.009493333597977957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,2560,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,12288,0.013572444518407186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,5120,0.026534222894244727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,3072,0.010435555544164447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,8192,0.01091733326514562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,16384,0.3610533343421088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,256,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,3584,0.03390844331847297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,7168,0.032255109813478254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,65536,0.10716178019841512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,12288,0.031823111904992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,10240,0.017857778403494094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,2048,0.03373333480623033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,2048,0.01757422255145179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,16384,0.10987022187974717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,6144,0.03434755404790243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,16384,0.027462222509913977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,32,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,2048,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,768,0.03012977706061469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,10240,0.016724444097942777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,3072,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,5120,0.010805333654085795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,8192,0.02629866699377696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,1024,0.0079644446571668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,4096,0.007724444071451823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,2560,0.007281777759393056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,2048,0.007027555671003129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,768,0.009397333694828881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,64,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,16384,0.014998222390810648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,128,0.025758221745491028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,2048,0.008656000097592672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,2048,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,128,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,768,0.003583999971548716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,65536,0.06552000178231134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,5120,0.10681511296166314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,12288,0.01882488860024346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,12288,0.013755555782053204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,128,0.005530666559934616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,6144,0.22973865932888457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,512,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,3584,0.017511111166742113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,1024,0.008252444366614023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,1536,0.0069146665434042616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,5120,0.02051555613676707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,128,0.012462221913867526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,1024,0.010732444624106089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,65536,0.2023022174835205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,128,0.0032151111712058387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,1024,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,3072,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,3584,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,7168,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,1536,0.007258666886223688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,128,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,2560,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,4096,0.01568800045384301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,5120,0.010320888625250923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,10240,0.01146133326821857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,512,0.00453333349691497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,10240,0.011535111400816174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,1536,0.02033244404527876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,3584,0.018274666534529794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,1024,0.003916444463862313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,12288,0.029271999994913738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,2560,0.013739555246300168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,128,0.014118222726715935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,2048,0.07359200053744845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,2560,0.028667555914984807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,1536,0.00869066682126787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,12288,0.12468266487121582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,16384,0.435169776280721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,3584,0.010000000397364298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,3072,0.007060444189442529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,7168,0.02383377816942003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,128,0.008727110922336578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,6144,0.012700444294346703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,768,0.02514755560292138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,7168,0.01292622254954444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,2560,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,3584,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,64,0.003921777837806278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,256,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,3072,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,16384,0.0775760014851888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,12288,0.02951822347111172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,10240,5.774355570475261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,64,0.01889333294497596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,1024,0.01501066651609209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,256,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,512,0.010483555495738983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,6144,0.11561955346001519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,10240,0.012102222277058495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,1024,0.012793777717484368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,10240,0.03125955661137899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,10240,0.025721778472264607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,2048,0.005558222118351195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,768,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,1536,0.0069004446268081665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,8192,0.010556444525718689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,3584,0.007653333246707916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,10240,0.39382134543524844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,128,0.013351110948456658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,256,0.003192000091075897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,12288,0.022778666681713525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,768,0.010227555202113258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,4096,0.01036800030204985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,65536,0.09063111411200629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,6144,0.03437866767247518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,7168,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,6144,0.013423110875818463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,4096,0.02477333280775282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,1536,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,256,0.011717333561844297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,1536,0.024079999989933435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,2048,0.00871377769443724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,65536,0.15708977646297878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,32,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,1536,0.016436444388495553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,10240,0.01516177753607432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,1536,0.006666666517655055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,768,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,4096,0.03939466675122579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,1024,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,2560,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,2560,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,256,0.003504000190231535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,1536,0.007017778025733099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,1536,0.03745955559942458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,1536,0.0377840002377828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,5120,0.008346666892369589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,16384,0.0831164452764723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,3072,0.01960000064637926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,8192,0.038809776306152344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,1536,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,1024,0.015476443701320224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,512,0.004847110973464118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,512,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,512,0.014544000228246054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,64,0.002814222127199173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,12288,0.01755288905567593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,3584,0.0075786668393347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,3584,0.015601777368121676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,128,0.014856000741322836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,1536,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,65536,0.06807466348012288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,64,0.0028275555620590844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,16384,0.02014755540423923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,4096,0.024101333485709295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,5120,0.008419555922349295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,4096,0.01588800052801768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,256,0.004233777936961916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,3584,0.020928889513015747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,4096,0.05159644285837809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,4096,0.028524445162879095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,4096,0.007901333272457123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,2560,0.0074417781498697065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,64,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,16384,0.022117333279715642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,8192,0.006682666639486949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,1024,0.019845333364274766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,65536,0.06368533107969496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,1536,0.006252444452709622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,1536,0.004254222330119875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,3584,0.017475555340449016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,32,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,6144,0.029232889413833618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,128,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,65536,3.1016328599717884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,7168,0.037984000311957464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,1536,0.11563289165496826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,16384,0.017599110802014668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,64,0.003162666741344664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,2048,0.01036444471942054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,2048,0.025295999315049913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,256,0.01512888901763492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,2560,0.007044444481531779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,2560,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,10240,0.011720000041855706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,10240,0.5075395372178819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,2048,0.007988444632954067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,1024,0.007669332954618666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,8192,0.04277955492337545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,512,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,4096,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,8192,0.027317333552572463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,1024,0.009476444787449306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,7168,0.022445332672860887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,2560,0.02942133280966017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,16384,0.08300977945327759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,32,0.003006222140457895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,1024,0.0042275554604000514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,8192,0.019314666589101154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,3584,0.007682667010360294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,8192,0.01611022154490153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,768,0.016508445143699646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,12288,0.02844711144765218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,512,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,2048,0.013441777891582914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,10240,0.01687466601530711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,3584,0.02411199940575494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,1024,0.013728000223636627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,1024,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,768,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,512,0.0038684445122877755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,2560,0.011758222348160215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,8192,0.025228444072935317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,1024,0.016579555140601266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,2560,0.022967111733224656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,6144,0.01831022236082289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,7168,0.033434665865368314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,5120,0.022641777992248535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,1024,0.005640888793600931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,768,0.025028443998760645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,256,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,65536,0.060043553511301674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,128,0.01483022173245748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,128,0.006239999913507038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,3072,0.008621333373917473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,5120,0.018187556001875136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,2560,0.026808000273174707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,768,0.016494222813182406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,12288,0.03196000059445699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,16384,0.06633422109815809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,16384,0.09823822312884861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,8192,0.16018933720058864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,128,0.014533332652515836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,2048,0.21554933653937447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,3584,0.012114666402339935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,12288,0.02437688906987508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,1536,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,16384,0.05299289027849833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,128,0.01811377704143524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,8192,0.04081511166360643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,7168,0.012373333175977072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,512,0.01481066644191742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,256,0.009714666340086196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,3584,0.012936000194814471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,2048,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,256,0.01928711103068458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,16384,0.043996443351109825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,7168,0.04680800106790331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,2048,0.011853333148691388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,4096,0.007976888782448238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,512,0.004179555508825514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,10240,0.012087999946541257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,7168,0.08947910865147908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,768,0.0052382221652401825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,128,0.016490666402710807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,10240,0.017625777257813346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,5120,0.013474666410022311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,10240,0.013071999781661563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,8192,0.16003110673692492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,10240,0.026368889543745253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,32,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,768,0.016210666961140104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,4096,0.07081155644522773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,2560,0.016534222496880423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,32,0.002830222249031067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,10240,0.9445288976033529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,65536,0.9433751636081271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,128,0.00452888881166776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,7168,0.03533066643608941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,5120,0.014821334017647637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,4096,0.011406222151385413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,1024,0.0041982221106688184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,4096,0.01754488878779941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,2560,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,3584,0.0141004439857271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,64,0.00833155545923445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,256,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,768,0.01443377799457974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,16384,0.04459289047453138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,2560,0.014523555835088095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,256,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,8192,0.026927111877335444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,6144,0.011521778172916837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,128,0.015778667396969266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,128,0.005666666560702854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,3072,0.019147555033365887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,64,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,7168,0.014489778214030795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,16384,0.016198222835858662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,8192,0.07037066751056247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,3072,0.026016889346970454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,65536,0.0738604466120402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,8192,0.02422488894727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,32,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,768,0.010839111275143094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,1024,0.014116444521480136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,5120,0.08167911238140531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,512,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,2560,0.015862221519152325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,2560,0.13212711281246609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,12288,0.008249777886602614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,1024,0.0063279999627007385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,8192,0.010363555616802638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,65536,0.06204266680611504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,64,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,65536,0.07360800107320149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,5120,0.008669333325492011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,16384,0.23208623462253145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,8192,0.05772799915737576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,2560,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,8192,0.011095999843544431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,6144,0.1255591048134698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,128,0.0029013332807355454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,1024,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,4096,0.015099555253982544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,6144,0.018906666172875296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,7168,0.06459555361005995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,10240,0.03673422336578369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,128,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,5120,0.041026665104760066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,6144,0.018057778477668762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,65536,0.1215333276324802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,65536,0.06467733118269178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,512,0.0032257777121331957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,2048,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,65536,0.15945333904690212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,2560,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,12288,0.19190666410658094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,2048,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,256,0.022368000613318548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,10240,0.06666844420962863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,256,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,7168,0.07457866933610705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,3072,0.07253155443403456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,512,0.006077333456940121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,65536,0.09051289161046346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,10240,0.01590488851070404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,10240,0.015747555428081088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,10240,0.025674665967623394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,512,0.018214222457673814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,2560,0.008303110798199972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,32,0.002855111120475663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,7168,0.01752800080511305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,7168,0.0638435549206204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,1024,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,3584,0.01762488815519545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,3584,0.020946666598320007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,2048,0.02949955397182041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,256,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,12288,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,10240,0.08201511038674249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,4096,0.05365955498483446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,12288,0.20590578185187447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,768,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,6144,0.03420533405409919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,1536,0.01670488880740272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,3072,0.008061333662933773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,4096,0.38332356346978086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,8192,0.019467555814319186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,256,0.024673778149816725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,7168,0.02125688890616099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,3584,0.016846223009957206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,6144,0.027313777142100867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,16384,0.0484444432788425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,65536,0.11907377507951523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,2560,0.008655110994974772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,256,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,256,0.007689778175618913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,4096,0.10587555832333034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,4096,0.010102222363154093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,65536,0.24285867479112414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,12288,0.024018666810459558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,64,0.007635555333561367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,1536,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,1024,0.014860444598727755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,64,0.005696889013051987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,2048,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,2048,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,768,0.005864000154866113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,1536,0.03458399905098809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,768,0.008616000413894653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,256,0.01239377756913503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,512,0.0041857775714662345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,512,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,12288,0.31957777341206867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,8192,0.057215107811821826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,3072,0.007972444097201029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,3584,0.020106666617923312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,8192,0.07252444161309136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,32,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,32,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,768,0.004580444345871608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,256,0.013068444199032254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,65536,0.0785013304816352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,4096,0.010415111151006486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,8192,0.010690666735172272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,2560,0.008626666333940294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,8192,0.01147377739350001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,3072,0.012585777375433179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,5120,0.009287111461162567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,2048,0.00665244460105896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,3584,0.011346666349305047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,12288,0.05854933129416572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,3584,0.009677333136399588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,2560,0.03641511003176371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,6144,0.040018667777379356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,16384,0.7998168733384875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,7168,0.04857066604826185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,32,0.002850666642189026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,3072,0.010072000324726105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,2048,0.015967110792795818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,1536,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,128,0.005197333378924264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,1536,0.004491555607981152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,2048,0.008813333180215623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,8192,0.011708444191349877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,6144,0.025984888275464375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,10240,0.012072000238630505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,64,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,3584,0.04281155599488152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,32,0.047636446025636464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,65536,0.06525244315465291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,10240,0.011928888658682505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,512,0.010239111052619087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,1024,0.01521511044767168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,768,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,2560,0.02781244450145298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,16384,0.02603911028967963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,256,0.003568888952334722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,3072,0.007275555696752336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,7168,0.01956622302532196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,3584,0.026741334133678015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,16384,0.027091556125217017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,5120,0.028703110085593328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,768,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,128,0.014474666780895658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,256,0.0042239998777707415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,3072,0.07111910978953044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,1536,0.02870044443342421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,10240,0.03730488816897074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,1024,0.00388355553150177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,1536,0.0069120000633928515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,256,0.017231111725171406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,1024,0.005987555616431766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,16384,0.03329688972897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,16384,0.07678933276070489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,65536,0.08576533529493545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,768,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,7168,0.6282435523139106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,768,0.010055111514197456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,8192,0.20513421959347197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,768,0.010392888552612728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,2048,0.011056000159846412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,128,0.023433778021070693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,32,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,3584,0.013615111509958902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,7168,0.033301333586374916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,1024,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,4096,0.008375999828179678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,16384,0.03488355543878343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,32,0.014816888504558139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,32,0.006593777901596493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,512,0.021967111362351313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,7168,0.020614221692085266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,12288,0.02612977723280589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,7168,0.012996444271670448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,10240,0.012753778033786349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,256,0.00564444437623024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,6144,0.03214755654335022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,3072,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,6144,0.02684266699684991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,32,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,1024,0.003903999924659729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,512,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,5120,0.008974221845467886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,64,0.003564444267087512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,1536,0.09265155262417263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,1024,0.01200888885392083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,4096,0.021984888447655573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,3072,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,4096,0.017314665847354464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,1024,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,6144,0.02164977788925171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,3584,0.0144533332851198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,2048,0.04253600041071574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,256,0.0041039999988343985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,256,0.023686221904224817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,1024,0.013771555489963956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,7168,0.006878222028414409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,128,0.011766222616036734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,8192,0.04252622193760342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,2560,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,7168,0.008172444171375698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,10240,0.2390142281850179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,1536,0.005521777603361342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,768,0.011787555283970304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,3072,0.0240657776594162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,3584,0.00657155571712388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,4096,0.007642666498819987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,5120,0.017232000827789307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,7168,0.17551022105746797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,64,0.002798222212327851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,128,0.014440889159838358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,7168,0.010034666293197209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,1536,0.00630666646692488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,5120,0.10411644644207424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,2560,0.007094221810499827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,3072,0.010262221925788457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,4096,0.022907555103302002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,10240,0.02792533238728841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,12288,0.013554666605260638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,3584,0.00865955568022198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,10240,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,10240,0.014825777875052558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,2560,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,2048,0.028251555230882432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,768,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,2560,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,6144,0.01755555636352963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,2560,0.01754133403301239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,65536,0.15901688734690347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,1536,0.013733333183659447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,5120,0.00906755526860555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,64,0.02490577763981289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,1024,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,65536,0.08275200261010064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,2560,0.03121333320935567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,1024,0.05259022116661072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,16384,0.024919999970330134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,16384,0.1325191126929389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,3072,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,2048,0.7788871129353842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,3584,0.016177778442700703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,1024,0.025723555021815833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,1024,0.0052044445441828836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,16384,0.0479928884241316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,5120,0.020218667056825426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,10240,0.019095111224386428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,768,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,768,0.015480889214409722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,3584,0.011522666447692446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,5120,0.031269331773122154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,1536,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,8192,0.03414488832155863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,64,0.0031306667046414483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,12288,0.18220444520314535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,3072,0.02202755543920729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,64,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,8192,0.02920088834232754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,3584,0.010678222609890832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,1024,0.017885333961910672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,8192,0.0229404436217414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,7168,0.006908444480763541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,768,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,12288,0.016511110795868766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,6144,0.14054489135742188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,2048,0.014599111345079211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,8192,0.08775911066267227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,32,0.006223111102978389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,4096,0.010315555665228102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,32,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,2048,0.014486221803559197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,3072,0.14460799429151747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,32,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,2560,0.014478221535682678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,8192,0.010767999622556897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,512,0.0068959999415609576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,7168,0.03994844357172648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,8192,0.018886221779717337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,512,0.00489066665371259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,128,0.003983111017280155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,4096,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,8192,0.011101333631409539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,8192,0.02299555473857456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,256,0.015119110544522604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,3584,0.009007111191749573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,128,0.014852444330851236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,65536,0.23968799908955893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,5120,0.03849688834614224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,128,0.01762488815519545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,256,0.010020444790522257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,128,0.007925333248244392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,12288,0.020765332712067496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,128,0.006924444602595435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,2560,0.016507556041081745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,3072,0.009215110705958473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,512,0.015507555670208402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,2048,0.010195555786291758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,256,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,16384,0.03909600112173293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,768,0.011909333368142446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,1536,0.014096000128322177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,8192,0.015835555063353646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,7168,0.031209776798884075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,3072,0.00885688927438524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,512,0.003540444291300244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,32,0.0029013332807355454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,64,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,65536,0.059608002503712974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,12288,2.1976356506347656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,2560,0.017680888374646504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,128,0.006974221931563483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,4096,0.020250666472646926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,2560,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,1024,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,3584,0.01128533316983117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,2560,0.007000889215204451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,16384,0.039573334985309176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,6144,0.030202666918436687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,768,0.004547555413511065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,3072,0.007360888852013483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,128,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,65536,0.04692000150680542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,6144,0.592724429236518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,2560,0.03654844562212626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,512,0.008246222303973304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,2560,0.033231112692091204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,64,0.012415111064910889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,12288,0.03976800044377645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,512,0.0034640000926123727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,2560,0.007260444263617198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,7168,0.011367999845080905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,3584,0.025840888420740765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,6144,0.00794044468137953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,1536,0.006620444357395172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,512,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,2560,0.02587022218439314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,7168,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,128,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,512,0.003612444632583194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,512,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,2560,0.013157332936922709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,128,0.034154666794670954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,128,0.04826577835612827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,256,0.2512977653079563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,256,0.016543999314308167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,1024,0.014475555883513557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,12288,0.14785688453250462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,256,0.0148062225845125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,6144,0.012412444584899478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,3584,0.037866665257347956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,6144,0.03306577934159173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,12288,0.0430026650428772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,768,0.006973333656787872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,3584,0.016480889585283067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,65536,0.05954933166503906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,128,0.011687999798191918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,5120,0.10109421941969131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,10240,0.019941333267423842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,1024,0.01622222198380364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,1024,0.015843555331230164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,32,0.026016889346970454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,10240,0.027375110321574744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,2048,0.5975182321336534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,5120,0.01588711142539978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,3072,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,3072,0.02363555630048116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,1536,0.012711111042234631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,64,0.004854222138722737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,65536,0.1317173375023736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,8192,0.042859554290771484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,4096,0.008634666601816813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,2048,0.006909333169460297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,1024,0.007621333003044128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,512,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,128,0.015079110860824585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,1536,0.010665777656767102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,5120,0.008613333106040955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,128,0.0032017777363459268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,64,0.0029013332807355454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,2560,0.009289777941173976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,3072,0.05466399921311272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,16384,0.021643555826610986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,16384,0.01695644524362352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,7168,0.025685333543353613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,8192,0.019945777124828763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,8192,0.02644266684850057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,1024,0.0038737778862317405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,256,0.010584889186753167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,65536,0.5106657875908746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,12288,0.02516711089346144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,512,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,8192,0.01091733326514562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,4096,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,2560,0.009912889036867354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,1536,0.004557333472702238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,10240,0.08721422486835056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,8192,0.0232595553000768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,512,0.006973333656787872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,256,0.010036444498433007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,10240,0.06751110818650988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,4096,0.014801777071423002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,5120,0.044734223021401294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,32,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,1024,0.07739466428756714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,1536,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,8192,0.007128888534175024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,2560,0.007113777928882175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,128,0.014448000325096978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,4096,0.04480977853139242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,16384,0.00794311116139094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,768,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,7168,0.07620622052086724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,4096,0.026564444104830425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,1536,0.012192000117566852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,3072,0.015094222293959724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,768,0.0035786665976047516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,4096,0.026397332549095154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,10240,0.03496355480617947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,2560,0.015615110596021017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,8192,0.01424088908566369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,10240,0.059614221254984535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,256,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,3584,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,256,0.05015733175807529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,65536,4.109844631618924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,8192,0.01977599991692437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,2560,0.013055110971132914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,16384,0.023723555935753718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,64,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,8192,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,32,0.007607111500369177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,1024,0.007632888853549957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,1536,0.006182222316662471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,2048,0.011029333704047732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,768,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,128,0.01483022173245748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,65536,0.025990222891171772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,1024,0.004875555634498596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,16384,0.04929066697756449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,4096,0.009924444059530893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,128,0.015196444259749519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,8192,0.01958933307064904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,2560,0.016498666670587327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,12288,0.01870755520131853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,128,0.0299928883711497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,32,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,65536,0.10499199893739487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,8192,0.024672889047198828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,65536,0.07243288887871636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,256,0.0032142222755485107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,3584,0.016140444411171805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,4096,0.007272889216740926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,65536,0.14228533373938665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,128,0.008061333662933773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,128,0.016179554992251925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,2560,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,10240,0.012113778127564324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,4096,0.023881778120994568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,3072,0.009324444664849175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,6144,0.013714666995737286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,256,0.026153778036435444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,10240,0.025422222084469263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,3072,0.07093066639370389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,32,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,512,0.3169564406077067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,3584,0.016352888610627916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,4096,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,768,0.008036444584528605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,3584,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,7168,0.011023999916182624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,3072,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,10240,0.027280888623661462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,1536,0.01075111081202825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,3072,0.006513777706358168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,6144,0.009307555854320526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,768,0.004882666799757215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,10240,0.015803555647532146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,256,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,256,0.009857777920034196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,7168,0.01440088947614034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,2560,0.01620533400111728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,128,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,512,0.017895999881956313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,10240,0.01623911162217458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,2560,0.17653866608937582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,3584,0.015211555692884656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,12288,0.010058666268984476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,512,0.0052897776994440295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,128,0.002875555513633622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,3072,0.01388533330625958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,3072,0.023708444502618577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,1024,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,5120,0.05988978015051948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,3584,0.04457599918047587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,32,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,128,0.005530666559934616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,4096,0.007681777907742395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,3072,0.014813333749771118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,3584,0.012104888757069906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,1024,0.009287999735938178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,2560,0.017832888497246634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,3072,0.012513777448071374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,6144,0.027679110566775005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,6144,0.010028444230556488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,5120,0.017039111918873258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,2560,0.01775733298725552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,2560,0.010675555302037133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,7168,0.04929955469237434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,1536,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,768,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,10240,0.012081777883900536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,3584,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,7168,0.0554257763756646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,2048,0.015489778584904142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,5120,0.009327111144860586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,65536,0.06358310911390516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,768,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,768,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,768,0.0041848888827694785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,32,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,128,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,512,0.0035031110876136353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,1536,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,7168,0.026718222432666357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,10240,0.05128444565667046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,65536,0.054367111788855664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,65536,0.06970489025115967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,2560,0.0076142218377855085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,512,0.013774221969975365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,3584,0.014544000228246054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,128,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,8192,0.021669333179791767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,3072,0.015536889433860779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,1536,0.025268443756633337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,2560,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,3072,0.012257777982287936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,6144,0.009717333647939894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,2560,0.009004444711738164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,3072,0.016332444217469957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,128,0.006165333506133821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,10240,0.04056088791953193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,1024,0.0039013334446483185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,3584,0.018750222192870248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,2560,0.006175111151403851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,2048,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,1536,0.010254222485754225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,256,0.014441778262456259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,64,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,12288,0.034352001216676496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,10240,0.011729777687125735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,1024,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,2048,0.015528000063366361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,1024,0.46376800537109375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,32,0.0035786665976047516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,3584,0.013150222599506378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,16384,0.13359999656677246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,65536,0.14081422487894693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,4096,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,1536,0.013184000220563678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,12288,0.027447111076778833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,1536,0.006779555645253923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,5120,0.020592000749376085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,7168,0.011339555184046427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,32,0.008216888540320927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,12288,0.013162666724787818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,12288,0.1042791141404046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,5120,0.012631999949614206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,1536,0.04014044337802463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,256,0.0035235554807715942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,10240,0.05554311143027412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,768,0.013229333692126803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,5120,0.014116444521480136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,3072,0.007911110917727152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,1536,0.006372444331645966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,1024,0.004975999808973736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,5120,0.020679111282030743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,65536,0.06004177861743503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,128,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,1536,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,768,0.014800889624489678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,12288,0.013282666603724161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,256,0.005698666804366642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,6144,0.016688888271649677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,512,0.022495110829671223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,1536,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,1024,0.014820444915029736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,7168,0.009648888475365108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,16384,0.041816890239715576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,5120,0.13234844472673205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,6144,0.029314666986465454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,4096,2.0231706831190324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,4096,0.015816888875431485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,1024,0.010651555326249866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,2560,0.006945778098371293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,2048,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,128,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,7168,0.00941244430012173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,5120,0.015471999843915304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,128,0.0075253330998950535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,10240,0.023967999551031325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,2048,0.01516177753607432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,12288,0.017072889539930556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,5120,0.01278222186697854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,3584,0.008267555799749162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,3584,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,65536,0.045703109767701894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,7168,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,6144,0.024634665913052026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,256,0.0031635556370019913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,3072,0.07239200009240045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,256,0.005797333187527127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,32,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,3072,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,768,0.02741777731312646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,4096,0.023183110687467787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,32,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,16384,0.020249777370029025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,12288,0.022069333328141108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,6144,0.010123555858929953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,512,0.024699555502997503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,32,0.0027751111321979095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,32,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,2560,0.008574222524960836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,5120,0.008296888735559251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,10240,0.057766218980153404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,512,0.02827022141880459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,768,0.017861333158281114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,5120,0.0683528913391961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,32,0.011747555600272285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,3072,1.0071662267049153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,3584,0.009830222361617619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,5120,0.17313244607713488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,12288,0.0380639996793535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,8192,0.016386666231685214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,128,0.30688977241516113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,2560,0.007092444433106317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,10240,0.01428533262676663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,768,0.005285333428117964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,1536,0.012507555385430654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,128,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,128,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,32,0.02047288914521535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,2560,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,2048,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,65536,0.2314382129245334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,16384,0.037647998995251127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,1536,0.018606222338146634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,16384,0.13283110989464655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,4096,0.03794666793611314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,2560,0.017852443787786696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,512,0.021648888786633808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,6144,0.03824888997607761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,7168,0.014418666561444601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,7168,0.01295199990272522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,8192,0.02053333322207133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,16384,0.02184533410602146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,512,0.0038240001433425476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,7168,0.04684355523851183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,3584,0.03773777683575948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,1024,0.1010284423828125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,10240,0.05269244313240051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,6144,0.01128355579243766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,10240,0.33460089895460343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,768,0.009703999592198266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,1024,0.02330400049686432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,512,0.00574222207069397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,768,0.00963288876745436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,4096,0.021653332644038733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,7168,0.015935111376974318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,65536,0.025441777375009324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,768,0.0036319999231232535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,3584,0.014767110347747803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,6144,0.009391110804345872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,1536,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,12288,0.04534133275349935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,6144,0.5308186742994521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,6144,0.033943110042148165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,3072,0.0127697777416971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,1024,0.02416977783044179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,3072,0.007607999775144789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,12288,0.019132445255915325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,768,0.005499555418888728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,2048,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,1024,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,2048,0.007248000138335758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,8192,0.023407111565272015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,16384,0.03194488750563727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,3072,0.03312622176276313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,7168,0.1608533329433865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,4096,0.01464355488618215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,7168,0.11012622382905747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,5120,0.1358284420437283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,2048,0.013879111243618859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,16384,0.010013333625263637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,3072,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,12288,0.033674667278925575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,6144,0.028684443897671167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,3584,0.27513599395751953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,256,0.0034835553831524322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,4096,0.019095111224386428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,7168,0.009691555466916826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,7168,0.021690666675567627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,1536,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,512,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,5120,0.012097777591811286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,3584,0.008101333346631791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,1536,0.006695999867386288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,2048,0.017165333032608032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,8192,0.018623999423450895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,256,0.016960889101028442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,512,0.005672889037264719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,6144,0.019343111250135634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,3072,0.017640888690948486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,4096,0.014480888843536377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,16384,0.07796177599165174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,4096,0.0922657781177097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,768,0.01734844512409634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,128,0.0616204473707411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,32,0.0029075555503368378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,1536,0.06342933575312297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,5120,0.01201600001917945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,3584,0.011416888899273343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,128,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,7168,0.05809511078728569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,7168,0.024504888388845656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,512,0.019596444235907663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,64,0.002830222249031067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,12288,0.014192000031471252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,3584,0.1076328886879815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,4096,0.008283555507659912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,8192,0.010749333434634738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,512,0.012303111453851065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,65536,0.0773493316438463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,3072,0.014175999495718213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,128,0.0041546668443414895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,16384,0.03561333484119839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,2048,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,7168,0.18657422065734863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,12288,0.02836444311671787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,3072,0.015126221709781222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,6144,0.029839111698998347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,4096,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,64,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,256,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,65536,0.07817511426077949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,3072,0.027144889036814373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,65536,0.08651110861036514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,16384,0.07945955461925931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,10240,0.012395555774370829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,12288,0.01348622226052814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,16384,0.5082897610134548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,4096,0.017162667380438913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,6144,0.008957333034939235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,3072,0.018535999788178336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,3584,0.008296888735559251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,65536,0.08613599671257867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,2048,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,32,0.020308444897333782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,4096,0.011374221907721626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,3072,0.007251555720965068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,16384,0.027113777067926195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,7168,0.02103022237618764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,768,0.005692444327804778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,65536,0.34310221672058105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,4096,0.011139555937714048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,256,0.005528888768619961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,6144,0.012057777908113269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,1536,0.008343111309740279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,12288,0.15142488479614258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,65536,1.1475448608398438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,32,0.016168889072206285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,65536,0.08370221985710992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,1536,0.01311644415060679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,8192,0.02986933456526862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,65536,0.23564089669121635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,1536,0.007996444072988298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,65536,0.060475554731157094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,7168,0.022069333328141108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,3584,0.08868000242445205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,10240,0.011767999993430244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,5120,0.014104000396198697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,128,0.013144888811641268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,65536,0.14356177382998997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,2048,0.03087288803524441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,256,0.024463999602529738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,2048,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,12288,0.4322204324934218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,7168,0.05744622151056925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,10240,0.2697128984663221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,256,0.011658666862381829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,4096,0.1487520005967882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,16384,0.2854115433163113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,2048,0.007304888632562425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,3584,0.01651733285850949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,5120,0.021706667211320665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,16384,0.0550871127181583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,3072,0.007419555551475949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,6144,0.012632889052232107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,10240,0.2545155684153239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,256,0.017870222528775532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,1536,0.008605333666006723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,10240,0.24139200316535103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,512,0.013201778133710226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,12288,0.03805422120624118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,768,0.006036444256703059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,3072,0.015140444040298462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,65536,0.02128355536195967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,5120,0.031340443425708346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,12288,0.017792888813548617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,512,0.027095110880004034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,7168,0.03986310958862305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,768,0.006219555520349079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,5120,0.015801777442296345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,3072,0.01624000072479248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,3584,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,768,0.0038622220357259116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,3584,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,10240,0.2374168766869439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,2048,0.13128356138865152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,2048,0.10108088784747654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,512,0.021919111410776775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,5120,0.008692444198661381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,256,0.015840000576443143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,1024,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,1536,0.009032888544930352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,5120,0.022275555464956496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,5120,0.01034577770365609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,12288,0.02365155518054962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,7168,0.00618577789929178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,512,0.006907555378145642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,7168,0.06624444325764973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,3584,0.011706666813956367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,10240,0.011778666741318174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,12288,0.03585244549645318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,5120,0.016211556063758004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,1536,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,16384,0.015503111812803479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,6144,0.01960444450378418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,7168,0.00996800015370051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,4096,0.043271111117468945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,3584,0.007974222302436829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,3072,0.0075777777367168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,3072,0.01278222186697854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,5120,0.018226666582955253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,4096,0.00924622184700436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,1536,0.0069271110826068455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,3584,0.09584889147016738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,2560,0.006569777925809224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,12288,0.019262222780121695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,128,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,256,0.004855999930037392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,2560,0.012749333348539142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,10240,0.011711999773979187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,12288,0.01313511116637124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,4096,0.0058853332367208265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,512,0.003560888684458203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,3584,0.014483556151390076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,2560,0.024668445189793903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,256,0.00997155573632982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,8192,0.014558222558763293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,2048,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,16384,0.008631111019187504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,256,0.0151582227812873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,512,0.0035128887328836652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,3584,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,2560,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,16384,0.012409778104888068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,128,0.004946666873163647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,7168,0.019666666785875957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,512,0.01422044469250573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,5120,0.012599111431174807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,4096,0.015398222539159985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,8192,0.012159999873903064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,1024,0.03792622354295518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,3584,2.1283270517985025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,3072,0.03195200032658047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,16384,0.023337778117921617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,3072,0.012098666694429187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,2560,0.006592000110281839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,512,0.005296000176005893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,6144,0.00902844468752543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,1536,0.004579555657174853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,6144,0.00904622260067198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,10240,0.007707555260923173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,768,0.06820799907048543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,2048,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,512,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,12288,0.01756533318095737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,256,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,6144,0.009642666412724389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,65536,0.209297776222229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,12288,0.023450666003757056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,65536,0.09380000167422825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,3072,0.008001777860853408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,3072,0.025808889004919264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,6144,0.05449155635303921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,12288,0.01143200033240848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,512,0.008176888856622908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,6144,0.011376889215575324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,16384,0.03719377848837111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,1024,0.016894222961531747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,256,0.022330666581789654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,5120,0.017151110702090792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,7168,0.01775733298725552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,10240,0.01347733371787601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,10240,0.006745777610275481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,65536,0.3197795550028483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,2560,0.11008977890014648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,5120,0.008676444490750631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,8192,0.014496889379289416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,2560,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,7168,0.026838221483760413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,1024,0.005998222364319696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,7168,0.2640106678009033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,10240,0.020231111182106864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,1536,0.006334222025341458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,64,0.002982222164670626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,2560,0.00599644457300504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,6144,0.4384320047166612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,768,0.016156444946924847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,10240,0.0704231129752265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,1536,0.03317777646912469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,12288,0.023038221730126276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,10240,0.01238222254647149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,2560,0.018460444278187223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,768,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,64,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,2560,0.008630221916569604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,6144,0.055063111914528735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,16384,0.041801777150895864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,1024,0.014463999205165438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,256,0.003621333175235324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,4096,0.01480888823668162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,128,0.003170666595300039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,2048,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,12288,0.017453332742055256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,128,0.011383111278216044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,5120,0.016565332810084026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,256,0.017224889662530687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,16384,0.0188755558596717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,2560,0.02576622201336755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,2048,0.024832889437675476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,2560,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,3584,0.007909333540333642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,3584,0.009323555562231276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,12288,0.01966399947802226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,32,0.0028159999185138275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,256,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,256,0.00369777778784434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,512,0.014512000812424554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,6144,0.012435555458068848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,3584,0.02102488941616482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,256,0.04505599869622124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,3584,0.00757244477669398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,4096,0.014656000667148165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,512,0.005574222240183089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,8192,0.057645334137810596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,64,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,3584,0.006263999889294307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,12288,0.014146667387750415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,8192,0.06688088840908475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,2048,0.048864000373416476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,4096,0.00830488900343577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,1536,0.018592000007629395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,3072,0.007896888587209914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,16384,0.0336444444126553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,3072,0.020950223008791607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,2048,0.12209422058529323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,5120,0.05726755327648587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,10240,0.015284443895022074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,768,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,16384,0.016161777906947665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,10240,0.03770044445991516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,10240,0.03254400028122796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,2048,0.008542222281297049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,1024,0.013798221945762634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,7168,0.018230222993426852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,32,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,10240,0.015842666228612263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,7168,0.05361066593064202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,256,0.0058648888435628675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,512,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,128,0.016223111086421542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,8192,0.011026666396194033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,4096,0.027765333652496338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,16384,0.027775999572541978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,65536,0.06879378027386136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,256,0.003847111016511917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,4096,0.009718221922715506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,8192,0.017880888448821176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,12288,0.016393777396943834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,16384,0.033997333712048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,512,0.012800888882742988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,128,0.01479555500878228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,5120,0.02267733381854163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,1536,0.02780533168050978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,2048,0.009350222018029954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,7168,0.009873777627944946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,4096,0.05239555570814344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,4096,0.03891466723548041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,6144,0.011993777420785693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,12288,0.12675021754370794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,1536,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,512,0.014044443766276041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,2560,0.026722666290071275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,768,0.022301332818137273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,2560,0.04169866773817274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,8192,0.023367111881573994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,8192,0.02363377809524536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,12288,0.12770488527086046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,4096,0.017961778574519686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,7168,0.018890667292806838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,8192,0.01514400045077006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,2048,0.03474400109714932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,1024,0.011875555747085147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,65536,0.10584266980489095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,10240,0.017287999391555786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,12288,0.024331554770469666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,128,0.0031226666437255014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,3072,0.01020088874631458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,12288,0.1657137738333808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,768,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,2560,0.020416888925764296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,2048,0.016513778103722464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,5120,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,1024,0.011890666352377998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,1024,0.0041680000722408295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,7168,0.11838489108615452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,5120,0.022104889154434204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,7168,0.013413333230548434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,768,0.016864889197879367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,2560,0.007000000112586551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,32,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,512,0.04389155573315091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,3072,0.015476443701320224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,16384,0.01586844523747762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,5120,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,7168,0.021316443880399067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,12288,0.029093331760830347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,5120,0.017539555827776592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,3584,0.007332444190979004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,10240,0.16769244935777453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,768,0.006673777682913675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,65536,0.06245511107974582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,10240,0.02829333477550083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,256,0.015228443675571017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,12288,0.03102133340305752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,1024,0.006187555690606435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,512,0.01648266613483429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,1536,0.013574221895800697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,4096,0.011412444214026133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,8192,0.013441777891582914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,512,0.01663822266790602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,512,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,2048,0.01978222197956509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,1024,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,32,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,7168,0.019498666127522785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,7168,0.007000000112586551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,512,0.0057760001056724125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,2560,0.07739377684063382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,10240,0.09305777814653184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,4096,0.008330666356616551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,2048,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,768,0.004181333300140169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,6144,0.018752889500723947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,512,0.008286221987671321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,1536,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,3584,0.026151999831199646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,7168,0.019783111082182992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,1536,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,8192,0.012423111332787408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,32,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,768,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,3584,0.007672000262472365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,256,0.015524443652894763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,8192,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,6144,0.013232889274756113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,2560,0.014886221951908536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,4096,0.014442667365074158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,6144,0.01815111107296414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,12288,0.01940622263484531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,256,0.004253333227501975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,4096,0.016493333710564505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,768,0.034612443712022566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,768,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,7168,0.5867013401455349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,8192,0.04555733336342705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,6144,0.017625777257813346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,5120,0.009680888719028896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,65536,0.12453777260250515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,32,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,2560,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,1536,0.008841777841250101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,3584,0.007975999679830339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,65536,0.13587110572391087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,256,0.0038684445122877755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,6144,1.0481671227349174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,5120,0.011643555429246692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,1536,0.02404266595840454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,1536,0.006173333360089197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,32,0.0033315554675128725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,128,0.009131555755933126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,8192,0.018560889694425795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,12288,0.020048000746303134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,4096,0.007668444679843054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,8192,0.08207199970881145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,1024,0.003871110992299186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,65536,0.08491466442743938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,6144,0.11250132984585232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,10240,0.027870221270455256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,7168,0.34096089998881024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,2560,0.021988444858127173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,768,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,12288,0.1091066665119595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,128,0.014819555812411837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,4096,0.012803555362754397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,2048,0.03849511014090644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,256,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,128,0.020909332566791113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,1536,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,16384,0.1681955522961087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,128,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,512,0.006197333335876465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,1024,0.01663644446267022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,12288,0.013923555612564087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,1536,0.004624888714816836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,1024,0.008027555214034187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,768,0.037017779217825994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,128,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,64,0.008309333688682979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,12288,0.023009777069091797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,6144,0.011397333608733283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,512,0.016898666818936665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,6144,0.027230223019917805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,2560,0.006933333145247565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,4096,0.012444444828563266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,3072,0.014161777165200977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,65536,0.07795200082990858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,4096,0.008984000318580205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,3072,0.010048888623714447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,3072,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,1024,0.020580444071027968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,4096,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,10240,0.08019733428955078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,2048,0.029137776957617864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,2048,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,768,0.006203555398517185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,3584,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,4096,0.00848355558183458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,65536,0.03033155534002516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,64,0.023167999254332647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,10240,0.01864888932969835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,10240,0.07790578073925443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,2560,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,256,0.013402666482660504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,3072,0.014107555150985718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,12288,0.03017866611480713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,2048,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,6144,0.3575502236684163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,2560,0.009723555710580613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,10240,0.022749332918061152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,2048,0.005517333332035277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,3584,0.03885066509246826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,3584,0.01889244384235806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,1536,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,5120,0.0407955547173818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,7168,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,8192,0.05678666962517632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,16384,0.02589599953757392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,512,0.0038355555799272326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,2560,0.018207111292415194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,4096,0.013418667018413544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,6144,0.012099555797047086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,512,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,16384,0.17462044292026094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,4096,0.12222311231825088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,3072,0.12733777364095053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,256,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,12288,0.026314665873845417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,10240,0.026544888814290363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,512,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,6144,0.00981422182586458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,3072,0.01940533353222741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,7168,0.03601866629388597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,1536,0.007782222496138678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,3584,0.023071999351183575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,512,0.014518222875065274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,3072,0.0075937774446275495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,10240,0.026350221700138513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,6144,0.0075733330514695905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,512,0.012076444096035428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,12288,0.07581244574652778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,3072,0.012808000048001608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,3584,0.018343999981880188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,32,0.0031475555151700974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,2048,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,32,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,768,0.013287111288971372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,4096,0.02071377800570594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,1024,0.013946667313575745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,12288,0.04274133178922865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,12288,0.00795555528667238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,3072,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,1024,0.0163128889269299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,1024,0.053935110569000244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,12288,0.01740622189309862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,3072,0.023031110564867657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,32,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,12288,0.042845331960254245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,10240,0.021571555071406897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,64,0.0058711109062035876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,3584,0.02370400064521366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,256,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,1024,0.05107466710938347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,5120,0.07514933082792494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,16384,0.04629688792758518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,12288,0.032771554258134626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,8192,0.046181334389580615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,65536,0.05537244346406725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,10240,0.03772266705830892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,2560,0.11266933547125922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,512,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,128,0.0041742221348815495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,8192,0.018778666853904724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,512,0.011397333608733283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,1536,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,64,0.002492444382773505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,1536,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,2560,0.009655999640623728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,3072,0.012764444781674279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,768,0.007736889024575551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,1024,0.003912000192536248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,3072,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,10240,0.01273155543539259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,512,0.013293333351612091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,64,0.012568888564904531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,7168,0.0776711106300354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,1536,0.024525332782003615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,1024,0.09256711271074082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,1536,0.029736889733208552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,3072,0.007672888537247975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,2048,0.017683555682500202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,16384,0.02996444371011522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,3584,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,10240,0.05572977993223402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,512,0.004920000003443824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,2560,0.025036444266637165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,8192,0.05123644404941135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,7168,0.006899555524190267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,7168,0.026400888959566753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,768,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,6144,0.011209777659840055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,5120,0.012083555261294046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,6144,0.013726222018400827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,128,0.0028159999185138275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,3584,0.023078221413824294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,10240,0.022484444909625586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,1024,0.014366222752465142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,512,0.06316088967853122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,5120,0.010684444672531553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,4096,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,10240,0.007899555895063613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,4096,0.00871377769443724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,12288,0.02349600030316247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,1536,0.026359111070632935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,256,0.0028880000528362063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,1024,0.003920000046491623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,12288,0.02365422248840332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,10240,0.015499555402331881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,2048,0.007003555695215861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,512,0.017504889104101393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,512,0.012252444194422828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,32,0.0041244443919923575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,3584,0.022527111901177302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,4096,0.010617777705192566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,3584,0.01364177796575758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,2560,0.2393857902950711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,4096,0.036271999279658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,1536,0.012000888586044312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,12288,0.04244622257020739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,65536,0.0914142198032803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,2048,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,10240,0.013767999907334646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,16384,0.08008088668187459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,512,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,1536,0.04002844293912252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,4096,0.06749955813090007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,12288,0.01684177749686771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,4096,0.029214223225911457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,6144,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,8192,0.030407110850016277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,4096,0.009689778089523315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,3072,0.03477866782082452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,8192,0.011404444773991903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,8192,0.13913155926598444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,10240,0.06517599688635932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,256,0.008606221940782335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,512,0.021355556117163763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,64,0.01383733335468504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,10240,0.015543111496501498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,64,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,256,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,6144,0.023054222265879314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,65536,0.08252355787489149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,12288,0.013136000268989138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,2560,0.04122311207983229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,3584,0.014448000325096978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,3072,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,768,0.01605599953068627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,7168,0.06740799877378675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,6144,0.04046133491728041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,65536,0.11710311306847467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,8192,0.01293333371480306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,32,0.005561777700980504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,6144,0.013748444616794586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,32,0.003141333245568805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,12288,0.02216177847650316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,256,0.004212444441186057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,256,0.010239111052619087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,8192,0.010811555716726514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,2560,0.015053333507643806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,1024,0.004943111290534337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,4096,0.010011555420027839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,2560,0.006978666616810693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,2560,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,3584,0.01074399964676963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,768,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,12288,1.6750657823350694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,64,0.014921777778201632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,65536,0.032355556885401406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,256,0.0052044445441828836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,128,0.027327999472618103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,65536,0.16594666904873318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,3584,0.011414222419261932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,64,0.0028524444335036804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,7168,0.03296711047490438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,12288,0.029143999020258587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,5120,0.014459555347760519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,1536,0.02176355487770504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,5120,0.01979733341270023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,768,0.0042160000238153665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,2048,0.01055466632048289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,5120,0.015923556354310777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,128,0.004977777600288391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,32,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,5120,0.025482666161325242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,2048,0.014449778530332776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,256,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,8192,0.2217031055026584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,1536,0.013808888693650564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,16384,0.05288799934917026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,2560,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,10240,0.013515555196338229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,5120,0.008665777742862701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,3072,0.15790043936835393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,768,0.018168000711335074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,2048,0.011950222154458364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,1536,0.006804444309737947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,3072,0.0063680000603199005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,3072,0.009378666679064432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,256,0.008640000389681922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,4096,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,8192,0.011904000408119626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,16384,0.033835556772020124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,256,0.02401155564520094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,65536,7.5538423326280375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,3584,0.07121421893437703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,2048,0.025806221697065566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,6144,0.058113780286577016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,1536,0.004954666727119022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,3584,0.012923555241690742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,32,0.011021333436171213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,10240,0.02549155553181966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,1536,0.006082666830884085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,1024,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,2048,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,2048,0.02565244502491421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,12288,0.01830400029818217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,4096,0.01144444445768992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,4096,0.015606222881211175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,3072,0.023589332898457844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,128,0.006609777609507243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,12288,0.04279199904865689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,512,0.005660444498062134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,3584,0.006899555524190267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,256,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,6144,0.019698666201697457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,2048,0.02529688841766781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,4096,0.013783999615245394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,8192,0.024127999941507976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,12288,0.03753511110941569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,5120,0.011040888726711273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,64,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,3072,0.011713777979214987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,768,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,32,0.0052177777720822235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,7168,0.02058844433890449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,7168,0.20105422867668998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,512,0.01720711092154185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,3584,0.006416000011894438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,768,0.006117333140638139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,3584,0.018206222189797294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,2048,0.012223111258612739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,1536,0.006656888872385025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,7168,0.011171555353535546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,1536,0.009399111072222391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,512,0.005752888818581899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,128,0.01579466627703773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,768,0.06488177511427137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,2048,0.020670221911536325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,1024,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,10240,0.03124888737996419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,8192,0.007191999918884701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,65536,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,512,0.015612444943851896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,6144,0.008700444466537898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,3072,0.01404888927936554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,16384,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,12288,0.0580879979663425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,512,0.01367377738157908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,65536,0.3808000087738037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,10240,0.05271822214126587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,768,0.03330755564901564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,4096,0.014532445205582513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,3584,0.01205244412024816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,5120,0.018416888184017606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,128,0.012049777640236748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,65536,0.09179377555847168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,2048,0.3951973385281033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,7168,0.027121777335802715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,6144,0.02808711263868544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,4096,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,768,0.003592888928121991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,7168,0.025063999825053748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,4096,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,128,0.012778666284349231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,3072,0.0069919998447100324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,2560,0.018760000665982563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,6144,0.018195556269751657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,2048,0.024683554967244465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,4096,0.014176888598336114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,7168,0.010229333407349056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,12288,0.01756088932355245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,768,0.014476444986131458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,4096,0.015984889533784654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,12288,0.027473777532577515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,6144,0.011301333705584208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,1024,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,512,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,128,0.00349866681628757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,128,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,10240,0.020901333954599168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,7168,0.03619022170702616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,2048,0.19191200203365752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,7168,0.030113776524861652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,32,0.006633777585294511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,1536,0.059996445973714195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,3072,0.007678222325113084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,65536,0.07502844598558214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,3584,0.00795199970404307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,8192,0.039510223600599505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,4096,0.014659555421935188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,128,0.003825777934657203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,16384,0.023344000180562336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,6144,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,128,0.0028328889360030494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,5120,0.010312000082598792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,12288,0.1594231128692627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,10240,0.018645332919226754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,3584,0.03568977779812283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,5120,0.008314666648705801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,8192,0.01718133356836107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,6144,0.08985066413879395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,12288,0.06338133414586385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,8192,0.02164622147878011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,3072,0.00794577764140235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,65536,0.2283057901594374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,3584,0.02632177703910404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,12288,0.019256000717480976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,16384,0.06415733363893297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,3072,0.083569778336419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,32,0.0028044444819291434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,4096,0.013011555704805585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,7168,0.01387466655837165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,10240,0.028003556860817805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,512,0.005968889014588461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,7168,0.026778666509522334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,7168,0.02362577782736884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,6144,0.009008888569143083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,65536,0.10115466515223186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,768,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,3584,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,1024,0.010239111052619087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,128,0.01313511116637124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,8192,0.014291556345091926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,2560,0.007243555453088548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,5120,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,2560,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,2560,0.013410666750537025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,7168,0.08362933662202622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,2048,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,5120,0.0151582227812873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,16384,0.5703271230061849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,3072,0.011415999796655444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,2560,0.0058355554938316345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,8192,0.2148808903164334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,2048,0.013863999810483722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,4096,0.007999999655617608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,1024,0.004308444344335132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,3072,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,6144,0.012721777790122561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,1536,0.03416266706254747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,16384,0.029768887493345473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,8192,0.26518132951524526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,10240,0.0705253349410163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,6144,0.020786666207843356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,2048,0.01830577850341797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,6144,0.01700622174474928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,1536,0.020599111914634705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,1024,0.00574933323595259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,6144,0.18710754977332222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,8192,0.016527111331621807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,2048,0.011694221860832639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,3584,0.013565333353148567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,3584,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,3584,0.061471111244625516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,7168,0.009727111293209923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,32,0.0025804444319672054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,65536,0.0532897777027554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,8192,0.02202488813135359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,65536,0.06265511115392049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,128,0.005668444352017508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,10240,0.02481066683928172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,1536,0.004582222137186262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,1536,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,5120,0.0188080006175571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,1024,0.009975999593734741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,1536,0.0292577776643965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,1024,0.6222151120503744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,1536,0.02077866593996684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,6144,0.03869955407248603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,512,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,6144,0.013199110825856527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,2560,0.01145333300034205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,10240,0.013085333009560903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,768,0.008461332983440822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,3072,0.07682489024268256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,512,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,65536,1.5967653062608507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,768,0.026081777281231348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,512,0.024697777297761705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,10240,0.015389333168665567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,2560,0.13718310991923013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,10240,0.055776887469821505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,3584,0.014492443866199918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,128,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,3584,0.011358222199810876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,3584,0.007648000286685095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,3072,0.007248889240953658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,512,0.011432888607184092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,512,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,3072,0.008057778080304464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,3072,0.018986667195955913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,65536,0.061780446105533175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,12288,0.021055110626750525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,5120,0.022533333963818018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,512,0.009343999955389235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,12288,0.04144622219933404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,10240,0.1801217794418335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,512,0.018031999468803406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,512,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,4096,0.01722577710946401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,12288,0.04328088959058126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,4096,0.02465244465404087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,16384,0.41830489370558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,2560,0.014534221755133735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,2560,0.014058666096793281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,128,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,512,0.003654222107595868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,3584,0.01645599967903561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,65536,0.055535111162397593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,16384,0.016210666961140104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,3072,0.015776889191733468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,256,0.01723377737734053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,5120,0.051783998807271324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,5120,0.012175111307038201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,1024,0.01512888901763492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,5120,0.010738666686746808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,4096,0.012239110966523489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,8192,0.031726221243540444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,2048,0.007982222570313348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,512,0.0033377777371141645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,4096,0.016515556308958266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,2560,0.05336266756057739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,8192,0.05696800020005968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,7168,0.020336889558368258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,1536,0.010054222411579555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,512,0.006620444357395172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,12288,0.023984000086784363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,768,0.005807111246718301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,3072,0.03186933199564616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,65536,0.014217777384652032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,7168,0.013145777914259167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,16384,0.03524355424775018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,6144,0.036566221051745944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,5120,0.009661333428488838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,65536,0.06312800116009183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,6144,0.019155555301242404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,10240,0.03285333183076646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,1536,0.0137155552705129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,65536,0.2589004569583469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,3584,0.007620444728268518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,4096,0.022835556003782485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,3072,0.014126222994592456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,16384,0.11539467175801595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,768,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,10240,0.10706666443083022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,2560,0.017835555805100333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,256,0.00527555536892679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,768,0.010738666686746808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,3584,0.007596444752481248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,16384,0.013212444053755866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,32,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,6144,0.018995554910765756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,6144,0.01961955593691932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,2560,0.021937777598698933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,128,0.020234665936893888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,512,0.00452977791428566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,128,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,128,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,3072,0.007287110719415877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,16384,0.0336426662074195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,768,0.015844444433848064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,512,0.05813866853713989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,12288,0.03084533413251241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,10240,0.01494488947921329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,256,0.0032239999208185407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,256,0.006672888994216919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,6144,0.009641778137948778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,4096,0.008010666403505538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,256,0.017599110802014668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,256,0.0031715554909573663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,3584,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,3584,0.007919111185603673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,768,0.009352888498041365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,65536,0.03998488850063748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,2560,0.006924444602595435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,7168,0.02236977716286977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,32,0.003145777723855443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,768,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,512,0.0068808889223469635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,12288,0.01793333391348521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,7168,0.013203555511103736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,768,0.008082666330867344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,512,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,65536,0.062089780966440834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,1024,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,5120,0.01715377800994449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,64,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,5120,0.008667555948098501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,10240,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,10240,0.037059555451075234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,5120,0.022255111071798537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,1536,0.030908445517222088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,16384,0.02608355548646715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,2048,0.011121778024567498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,2048,0.009718221922715506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,8192,0.2281280093722873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,2560,0.006639110959238476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,6144,0.01035911093155543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,8192,0.09253066778182983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,5120,0.009311111436949836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,8192,0.07501333289676242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,128,0.018219555417696636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,4096,0.033407999409569636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,2048,0.035931554105546736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,768,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,5120,0.20099377632141113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,3584,0.008032889001899296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,768,0.0038684445122877755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,1536,0.01145333300034205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,4096,0.04049955474005805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,512,0.003903111235962974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,3584,0.015528000063366361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,64,0.0032186667538351486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,512,0.01757244434621599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,6144,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,3584,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,10240,0.025428444147109985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,768,0.01551999979548984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,2048,0.0075839997993575195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,5120,0.02444000045458476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,4096,0.033626665671666466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,16384,0.05302933189604017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,32,0.004226666771703296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,8192,0.014504000544548035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,1024,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,4096,0.01641244524055057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,16384,0.2409342130025228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,3072,0.007497777541478475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,4096,0.015969778100649517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,10240,0.026149332523345947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,5120,0.022657778528001573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,2560,0.01036444471942054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,512,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,7168,0.02956977817747328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,128,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,512,0.014516444669829475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,2560,0.007276444799370236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,256,0.015480889214409722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,512,0.005316444569163852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,128,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,3584,0.014554666148291694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,12288,0.018183110488785636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,7168,0.14943822224934897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,10240,0.015708444847000968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,128,0.005538666827811136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,2560,0.009686222506894005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,2560,0.02201066745652093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,1024,0.003911111089918348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,10240,0.024093333217832778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,1024,0.003858666867017746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,6144,0.018195556269751657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,768,0.021783111823929682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,4096,0.010838222172525195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,2048,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,3584,0.00796622203456031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,3072,0.005662222289376789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,128,0.006320888797442119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,256,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,256,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,32,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,7168,0.03150755498144362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,512,0.016862221890025668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,3584,0.023222222924232483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,512,0.0031866667171319327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,1536,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,2560,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,2560,0.030118223693635728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,1536,0.009496889180607265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,128,0.006353777729802662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,3584,0.012306667036480375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,7168,0.09336355659696792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,768,0.004921777794758479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,32,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,12288,0.031567109955681696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,1536,0.02197155521975623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,2048,0.012094222009181976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,16384,0.023330666952663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,6144,0.05682400200102064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,6144,0.010012444522645738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,65536,0.06658666663699679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,6144,1.4279004202948675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,1536,0.011733333269755045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,65536,0.08110311296251085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,4096,0.012421333127551608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,65536,0.10836266809039646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,5120,0.018519111805491976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,6144,0.008646222452322641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,8192,0.022709333234363135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,512,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,3072,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,128,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,128,0.0030871110243929755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,12288,0.23629244168599448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,3072,0.029162665208180744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,3584,0.017966222431924608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,16384,0.31271110640631783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,12288,0.19051378303103975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,7168,0.010039110978444418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,10240,0.04333777891265022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,8192,0.014305777019924588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,8192,0.03876533442073398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,10240,0.010416888528399997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,3072,0.13423111703660753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,1024,0.022055110997623865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,10240,0.012080888781282636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,64,0.0028488888508743713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,1024,0.013433777623706393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,8192,0.09984355502658421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,65536,0.0721671117676629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,16384,1.8132461971706813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,1536,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,1024,0.019342222147517733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,2048,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,10240,0.17147821850246855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,4096,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,65536,0.19703377617730033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,3072,0.007293333609898885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,5120,0.0140328887436125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,3072,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,128,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,128,0.00453599997692638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,10240,0.016386666231685214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,10240,0.007293333609898885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,512,0.013451555536852943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,10240,0.014767999450365702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,128,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,128,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,3072,0.006385777973466449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,65536,0.1408240000406901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,256,0.003206222214632564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,1024,0.01551999979548984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,8192,0.06384622388415866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,512,0.011785777906576792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,2048,0.006568000134494569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,16384,0.021373333202468023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,256,0.0070293330483966405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,3584,0.017916444275114272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,1536,0.01585244470172458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,12288,0.03636622097757127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,3072,0.02090222140153249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,10240,0.021659556362364028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,256,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,128,0.0038515557017591265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,3072,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,768,0.007624000310897827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,3072,0.017443555924627516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,6144,0.009638222555319468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,128,0.015124445160230001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,6144,0.055352889829211764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,10240,0.016866667403115165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,128,0.02308533257908291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,128,0.017200888858901132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,16384,0.016843555702103507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,512,0.05538577834765116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,7168,0.02682222260369195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,16384,0.038042667839262224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,3584,0.013152889079517789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,3072,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,2048,0.022978666755888198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,1536,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,256,0.0032239999208185407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,6144,0.017234666479958426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,4096,0.028743998871909246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,1024,0.00979288915793101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,12288,0.022568888134426538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,10240,0.026162667406929865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,6144,0.01902755598227183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,128,0.0028408887899584244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,2048,0.006572444405820634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,32,0.008612444831265343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,3072,0.007312888900438945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,16384,0.27159733242458767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,8192,0.14785777197943792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,4096,0.009917333722114563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,128,0.0052791109515561005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,16384,0.045689778195487134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,512,0.014085332552591959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,2048,0.010363555616802638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,12288,0.034415112601386175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,8192,0.013930666777822705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,256,0.012804444465372296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,16384,0.06715022193060981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,8192,0.013506666653686099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,16384,0.030563556485705908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,5120,0.025608888930744592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,1536,0.009714666340086196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,6144,0.02377333409256405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,12288,0.023248889380031165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,256,0.014780445231331719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,5120,0.008655110994974772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,4096,0.008086221913496653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,12288,0.013478221992651621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,8192,0.027544000082545813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,5120,0.016862221890025668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,64,0.0027928888383838865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,2560,0.08080622222688463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,128,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,1536,0.01447111037042406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,65536,0.09411378039254083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,5120,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,8192,0.009007111191749573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,256,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,65536,0.1487857765621609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,768,0.016960889101028442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,16384,0.03182222114668952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,10240,0.018186666899257235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,1024,0.06996177964740329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,16384,0.05521777934498257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,3072,0.024901333782407973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,128,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,3072,0.032511999209721885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,2560,0.014116444521480136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,3072,0.01797155539194743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,8192,0.026191110412279766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,16384,0.04127733243836297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,12288,0.5363048977322048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,3072,0.015944888194402058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,1536,0.012091555529170565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,2048,0.08843644460042317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,256,0.011742221812407175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,65536,0.13321867254045275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,768,0.028373334142896865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,6144,0.012825777961148156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,5120,0.07465510898166232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,10240,0.06085599793328179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,5120,0.01090666651725769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,128,0.006592000110281839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,65536,2.1853983137342667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,3584,0.02127911150455475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,3584,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,65536,0.12605067094167074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,768,0.0042035554846127825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,10240,0.015491555134455362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,8192,0.031960888041390315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,65536,0.061089780595567494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,256,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,128,0.005727111051479976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,16384,0.021306667062971327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,12288,0.02307644486427307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,4096,0.014328888720936246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,768,0.029271112547980413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,3072,0.017879999346203275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,1024,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,2560,0.02346222268210517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,6144,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,768,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,65536,0.06199377775192261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,64,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,512,0.0035093331502543557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,1024,0.00795199970404307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,128,0.003536888708670934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,3584,0.022613333331214056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,5120,0.01071466671095954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,8192,0.024089778463045757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,768,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,8192,0.01958755486541324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,3584,0.025036444266637165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,16384,0.045069333579805165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,32,0.003162666741344664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,10240,0.014695111248228284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,1536,0.012534221841229333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,7168,0.012096000214417776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,3072,0.02204888893498315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,256,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,65536,0.06253155734803942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,6144,0.03501866592301263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,3072,0.028189334604475234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,12288,0.031126221021016438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,65536,0.08246044317881267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,10240,0.018424888451894123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,768,0.01684888866212633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,2048,0.011026666396194033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,512,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,8192,0.1565155585606893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,2048,0.015266666809717814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,1536,0.012697777814335294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,2048,0.014508444401952954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,512,0.00902133352226681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,3072,0.015895111693276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,1024,0.0042257776690853965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,2048,0.019967110620604623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,10240,0.019907555646366544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,65536,0.02851466668976678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,3072,0.015198222464985318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,128,0.013402666482660504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,3072,0.007310222420427535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,6144,0.00997511131895913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,8192,0.020248888267411124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,16384,0.10421688689125909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,2048,0.015462223026487561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,6144,0.023751111494170293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,8192,0.019274666905403137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,4096,0.01053066634469562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,12288,6.3126924302842875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,512,0.05579555696911282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,10240,0.029612445169025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,6144,0.04395555456479391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,65536,0.08002577887641059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,16384,0.03641066617435879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,65536,0.07522577709621854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,65536,0.07921866575876872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,12288,0.018363555272420246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,3072,0.037096887826919556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,128,0.01479200025399526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,3072,0.1398844454023573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,1536,0.04908711049291822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,12288,0.017215111189418368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,2048,0.06794310940636529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,2560,0.029682666063308716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,2048,0.006502222269773483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,6144,0.01754577789041731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,65536,0.11808266904619004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,128,0.005251555393139522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,256,0.004265777766704559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,2560,0.014473777678277759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,7168,0.11249511771731907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,1024,0.022084444761276245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,512,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,768,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,6144,0.027229333917299908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,3584,0.01443733274936676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,1024,0.01149511088927587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,6144,0.011016888750924004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,1024,0.01310755560795466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,2560,0.02442577812406752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,1024,0.012429333395428128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,7168,0.01886488828394148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,3072,0.008339555727110969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,4096,0.120871106783549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,512,0.006170666466156642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,2048,0.006212444355090459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,7168,0.015224888920783997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,2048,0.02201866606871287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,512,0.006606222026877933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,512,0.01817244456874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,2560,0.017736888594097562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,12288,0.014089778065681458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,3584,0.08172711398866442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,8192,0.02311022248533037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,8192,0.010999999940395355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,512,0.022632888621754114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,4096,0.23250222206115723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,12288,0.049280888504452176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,6144,0.03078311019473606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,3584,0.034296890099843345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,6144,0.023489778240521748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,8192,0.010438222024175854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,65536,0.07763733466466267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,1536,0.04153777824507819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,16384,0.0435751113626692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,12288,0.02104266650146908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,8192,0.07936622036827935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,256,0.015662221444977652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,2560,0.025773333178626165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,512,0.014862222803963555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,1024,0.019242667489581637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,2560,0.01754577789041731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,3072,0.007664889097213745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,16384,0.019820445113711886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,512,0.015812445018026564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,256,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,5120,0.03703377644220988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,2560,0.01423111061255137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,16384,0.013083555632167392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,256,0.004249777644872665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,12288,0.417813327577379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,128,0.00629333323902554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,32,0.0032213332338465583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,16384,0.07498577568266127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,5120,0.011361777782440186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,2560,0.03284444411595663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,64,0.004583999928500918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,128,0.003809777812825309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,65536,0.06396444638570149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,3584,0.007616889145639207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,7168,0.019280888968043856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,1536,0.0045422220395671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,64,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,64,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,128,0.013756444056828817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,8192,0.00833155545923445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,128,0.009947555760542551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,4096,0.09560889005661011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,256,0.017135999268955655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,6144,0.021436444587177698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,6144,0.009103111094898647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,3072,0.03166755702760484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,6144,0.011229332950380115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,128,0.025232000483406916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,7168,0.03951377669970194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,5120,0.016527111331621807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,1536,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,16384,0.015879111157523263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,32,0.03172711200184292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,3584,0.07777511411243014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,1024,0.004249777644872665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,5120,0.026949332820044622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,8192,0.1977848874198066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,4096,0.027309333284695942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,2048,0.006455111006895701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,128,0.02637333340115017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,256,0.01719377769364251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,3072,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,2560,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,8192,0.04496622085571289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,256,0.33046311802334255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,1024,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,768,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,8192,0.015617777903874716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,5120,0.009999111294746399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,5120,0.013728888498412238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,2560,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,64,0.03741066654523214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,256,0.005249777601824866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,256,0.019163555569118924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,6144,0.011386666860845355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,8192,0.13138400183783636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,65536,0.2651200029585096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,65536,0.23191910319858125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,128,0.014840889308187695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,3072,0.020222221811612446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,65536,0.06546844376458062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,2048,0.04604444570011563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,16384,0.020607110526826646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,2048,0.012322666744391123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,5120,0.0185253338681327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,128,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,3072,0.007439110842016008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,12288,0.03286044465170966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,12288,0.008173333273993598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,1536,0.00832799987660514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,1024,0.0038328886859946777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,1536,0.01480888823668162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,256,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,16384,0.04327199856440226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,12288,0.011331555744012197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,128,0.012053333222866058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,7168,0.019567999574873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,512,0.03838222225507101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,16384,0.013479111095269522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,4096,0.013595555391576556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,768,0.006319111006127463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,3584,0.012121777567598553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,4096,0.017916444275114272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,1024,0.006335999816656113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,2560,0.007685333490371704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,1024,0.013361777696344586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,2560,0.01015377789735794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,1024,0.003917333152559069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,4096,0.05123022198677063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,6144,0.04253777861595154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,128,0.01311466677321328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,768,0.02314044369591607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,512,0.0035146665241983202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,128,0.015447111593352424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,65536,0.1964906718995836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,16384,0.024653333756658766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,10240,0.025057777762413025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,6144,0.045422222879197865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,32,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,2048,0.017097777790493436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,2048,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,512,0.00407644444041782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,512,0.008311111066076491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,10240,0.012670222255918713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,512,0.016543999314308167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,3584,0.0075777777367168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,256,0.0041893331540955436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,512,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,10240,0.013530666629473368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,3072,0.00868444475862715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,128,0.01688799924320645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,16384,0.05505866805712382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,768,0.027104889353116352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,64,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,2048,0.015533333023389181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,12288,0.032391998502943255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,768,0.006252444452709622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,256,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,768,0.0063440000845326324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,256,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,768,0.0069368887278768755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,65536,0.10258933570649888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,12288,0.25909601317511666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,2560,0.00999644481473499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,512,0.0038195554580953387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,12288,0.021184888150956895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,3072,0.00657155571712388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,7168,0.012853333519564735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,3072,0.021145777569876775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,512,0.012092444631788464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,4096,0.04591822293069628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,8192,0.0451546675629086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,5120,0.02895555562443203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,1024,0.037010666396882795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,2048,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,6144,0.0207013338804245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,3584,0.0473217765490214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,1536,0.004573333180612988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,6144,0.01275199982855055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,65536,0.20757955975002715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,3072,0.015665777855449252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,2560,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,3584,0.008135110967689091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,10240,0.14996177620357937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,10240,0.020582222276263766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,128,0.0052791109515561005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,256,0.01295199990272522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,12288,0.017581333716710407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,2560,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,65536,0.14339733123779297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,5120,0.011868444581826528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,7168,0.021713778376579285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,768,0.007366221812036302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,512,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,2560,0.007150222029950883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,1024,0.008619555168681675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,3072,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,16384,0.013375110924243927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,768,0.008278222547637092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,7168,0.014504889647165934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,2560,0.0957991083463033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,8192,0.01073066641887029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,8192,0.01867733399073283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,1024,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,2048,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,8192,0.030245333909988403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,4096,0.7767190933227539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,256,0.01479288935661316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,64,0.009526222116417354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,3072,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,16384,0.4988382127549913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,8192,0.029861334297392104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,4096,0.007610666255156199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,4096,0.03046222196684943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,2048,0.015062222878138224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,512,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,12288,0.01992177797688378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,2048,0.01178311142656538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,2048,0.018199111024538677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,12288,0.016528889536857605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,64,0.0052391112678580815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,1536,0.07189422183566622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,6144,0.013463999662134381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,3584,0.05585155884424845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,8192,0.041068444649378456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,10240,0.13519022199842665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,4096,0.035170667701297335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,1024,0.012072000238630505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,512,0.005687110953860813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,6144,0.009544000029563904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,7168,0.013050666285885705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,5120,0.010392888552612728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,5120,0.009977777798970541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,1536,0.11951111422644721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,4096,0.01625866691271464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,32,0.008095111283991072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,8192,0.01408266690042284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,2048,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,3584,0.016528000434239704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,512,0.011364444262451597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,5120,0.008971555365456475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,7168,0.019291554888089497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,2560,0.012455999851226807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,64,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,1536,0.0163102216190762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,12288,0.3695733282301161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,8192,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,1536,0.0057706667317284485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,1536,0.008011555506123437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,6144,0.0947635571161906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,6144,0.008975110948085785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,8192,0.016176889340082806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,6144,0.013772444592581855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,1024,0.008050666915045844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,8192,0.01184977756606208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,5120,0.03644444545110067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,32,0.0028239999794297745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,64,0.02718488872051239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,2048,0.006751110984219446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,2048,0.009287111461162567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,256,0.017160000072585214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,65536,0.18065422111087373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,2048,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,1024,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,8192,0.3752266565958659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,512,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,8192,0.014123555686738757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,16384,0.12108622656928168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,2048,0.11482044061024983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,1024,0.007948444121413762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,10240,0.15429511335160997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,2048,0.02609066665172577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,2048,0.04590222239494324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,256,0.011726222104496427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,32,0.0031635556370019913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,1024,0.00964355551534229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,768,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,128,0.003528000166018804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,7168,0.01274222218328052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,256,0.015127999915017022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,1024,0.007511110769377814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,256,0.008648888932334052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,10240,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,6144,0.013456000222100152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,7168,0.013009777499569787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,32,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,65536,2.851050694783529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,10240,0.018952000472280715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,7168,0.054089777999454074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,64,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,16384,0.03184444374508328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,512,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,64,0.003173333282272021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,12288,0.017014222012625802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,10240,0.11348888609144424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,16384,0.04889511068662008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,4096,0.10175733433829413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,768,0.01753777762254079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,12288,6.188737657335069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,64,0.009275555610656738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,3584,0.028078221612506445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,65536,6.524359809027778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,768,0.004558222161398994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,2048,0.012434666355450949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,8192,0.010688888529936472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,16384,0.02661866611904568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,3584,0.01626755628320906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,16384,8.338053385416666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,5120,0.02332533399264018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,3584,0.029198222690158423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,512,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,4096,0.008354666332403818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,65536,0.10079733530680339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,512,0.011419555379284753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,512,0.004233777936961916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,6144,0.018228444788191054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,16384,0.5991804334852431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,128,0.015140444040298462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,1024,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,5120,0.042909334103266396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,6144,0.12473244137234157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,2560,0.009296889106432596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,4096,0.019526221685939364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,3584,0.008974221845467886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,128,0.005510222166776657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,2048,0.008441777692900764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,2560,0.19193689028422037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,4096,0.007601777712504069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,1536,0.004864888886610667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,65536,0.06301155355241564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,1024,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,2048,0.00685422205262714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,5120,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,7168,0.011932444241311816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,4096,0.00775111135509279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,6144,0.009658666948477427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,12288,0.16260622607337102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,65536,0.06692088974846734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,10240,0.02130222154988183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,1024,0.026683555708991155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,6144,0.032919999625947736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,3584,0.016918222109476726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,10240,0.10061689217885335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,3072,0.007963555554548899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,65536,0.06266755527920194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,2048,0.014128888646761576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,16384,0.041962666643990405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,6144,0.027926223145590887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,6144,0.049350221951802574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,128,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,10240,0.015501333607567681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,8192,0.13687288761138916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,2560,0.025832000705930922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,512,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,5120,0.00832622249921163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,12288,0.10074133343166775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,4096,0.00849955528974533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,2560,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,10240,0.01682133310370975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,128,0.038231998682022095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,64,0.0031600000543726813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,512,0.003504000190231535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,5120,0.009015999734401703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,12288,0.3528444502088759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,512,0.0034906665484110513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,8192,0.03143999973932902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,2048,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,1024,0.006252444452709622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,2048,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,2048,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,5120,0.008969777988062965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,2048,0.015263110399246216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,65536,1.1020142237345378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,7168,0.015294222368134392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,3072,0.014664888381958008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,5120,0.03050133254792955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,128,0.004954666727119022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,256,0.012406222522258759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,3072,0.009732444253232744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,65536,0.07670844263500638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,5120,0.024056888288921777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,3584,0.007366221812036302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,768,0.017885333961910672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,3072,0.02008800043000115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,3584,0.013961777091026306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,1024,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,6144,0.1503208875656128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,65536,0.113155550426907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,768,0.014160888062583076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,64,0.0028000000036425064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,256,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,4096,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,768,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,3584,0.031081779135598078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,1536,0.010759111079904767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,12288,0.02570844524436527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,3584,0.008367111285527548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,2560,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,1536,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,256,0.0042088888585567474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,8192,0.08903466992908055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,2048,0.04058577617009481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,8192,0.010370666782061258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,1536,0.016540444559521146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,12288,0.017271111408869427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,10240,0.0099982221921285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,12288,0.5163991186353896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,65536,0.10887910922368367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,2048,0.00674311113026407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,1024,0.009096889032257928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,768,0.014123555686738757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,3072,0.013012444807423485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,65536,0.06260710954666138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,4096,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,3584,0.01593155496650272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,128,0.01478488908873664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,10240,0.018512000640233357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,5120,0.026368889543745253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,256,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,512,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,8192,0.010374222364690568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,768,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,7168,0.01887999971707662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,64,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,256,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,7168,0.010037333601050908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,10240,0.013804444836245643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,256,0.03572533196873135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,2560,0.011740444435013665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,2560,0.052277333206600614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,512,0.006258666515350342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,3584,0.028031110763549805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,16384,0.1665875514348348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,3072,0.00832177781396442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,6144,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,64,0.003547555456558863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,8192,0.0352862212393019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,65536,0.06397333410051134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,8192,0.01348888874053955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,3584,0.029250668154822454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,2560,0.023808000816239253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,2560,0.013735999663670858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,8192,0.03152889013290405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,16384,0.04034311241573758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,512,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,1024,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,1536,0.016308445069524977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,5120,0.008279110822412703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,7168,0.01344622257683012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,128,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,2048,0.006873777757088344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,16384,0.01990133358372582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,1536,0.00517155561182234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,3072,0.023583999938435022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,1536,0.00453333349691497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,256,0.033077332708570696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,7168,0.01258488910065757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,768,0.00655733338660664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,512,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,6144,0.018509333332379658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,512,0.0035413333939181436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,2048,0.03410133388307359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,768,0.015489778584904142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,256,0.005990222096443176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,7168,0.1277279986275567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,65536,0.09207377831141154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,3072,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,4096,0.030219554901123047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,4096,0.007999999655617608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,768,0.020560888780487906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,5120,0.016148444679048326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,2560,0.00796088907453749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,128,0.005215999980767568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,7168,0.014292443792025248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,2048,0.014072888427310519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,12288,0.06791555881500244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,768,0.010661332971519895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,3584,0.00906044410334693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,128,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,7168,0.020647999313142564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,3072,0.014999111493428549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,1536,0.0069919998447100324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,2048,0.02347999976740943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,128,0.02204799983236525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,7168,0.06822666856977674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,3584,0.007603555917739868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,4096,0.019197333190176222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,16384,0.016145777371194627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,2048,0.00654755574133661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,16384,0.039523555172814265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,4096,0.00794222205877304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,65536,0.0654675563176473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,512,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,256,0.015152000718646579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,2048,0.009278222090668148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,16384,0.021212443709373474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,8192,0.01591822173860338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,12288,0.043960889180501304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,32,0.1675128936767578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,16384,0.10677244265874226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,1024,0.03369244601991441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,64,0.009418666362762451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,12288,0.18285955323113334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,2048,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,65536,4.0246632893880205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,4096,0.012539555629094442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,16384,0.3072613345252143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,128,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,64,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,256,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,7168,0.010247999595271217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,7168,0.010669333239396414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,3584,0.011153777440388998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,32,0.0032675556010670135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,16384,0.05773155556784736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,7168,0.012870222330093384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,1024,0.006156444549560547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,128,0.005680888891220093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,256,0.015227556228637695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,2560,0.01610577768749661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,7168,0.01421866648726993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,512,0.007617777420414819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,5120,0.027090667022599116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,3072,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,128,0.013077333569526672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,5120,0.03173866536882188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,16384,0.165175994237264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,32,0.004594666676388847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,12288,0.05425777700212267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,10240,0.04104355639881558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,1024,0.005107555538415909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,7168,0.014797333214018079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,3072,0.04559288753403557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,2048,0.02436711059676276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,2048,0.02500533395343357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,512,0.0035253332720862497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,2560,0.010406222608354356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,256,0.0147706667582194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,8192,0.014866666661368476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,128,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,768,0.0038444445365005066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,5120,0.09340177641974555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,4096,0.09720444679260254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,6144,0.023288889063729182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,768,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,8192,0.014766222900814481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,7168,0.015524443652894763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,65536,0.046718223227394946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,768,0.015650666422314115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,12288,0.018192888961897958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,512,0.0032142222755485107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,512,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,5120,0.0099982221921285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,10240,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,4096,0.013799111048380533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,12288,0.02093066606256697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,10240,0.020587555236286588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,5120,0.03858222232924567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,5120,0.009551111194822524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,4096,0.02030044462945726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,2048,0.02334311107794444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,6144,0.010838222172525195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,768,0.004538666870858935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,7168,0.014728888869285583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,1024,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,128,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,7168,0.08742666906780666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,128,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,65536,0.18317333857218424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,3072,0.00998044427898195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,10240,0.05714222457673815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,6144,0.009372444616423713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,10240,0.015051555302408008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,7168,0.03260622090763516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,4096,0.022991999983787537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,64,0.02275377843115065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,64,0.002798222212327851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,65536,0.07058311171001859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,10240,0.03023911184734768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,256,0.03033422099219428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,3072,0.03133777777353922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,12288,0.24061687787373862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,512,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,1536,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,12288,0.013232000172138214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,2048,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,5120,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,512,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,7168,0.49163468678792316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,1536,0.012323555847009024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,32,0.004582222137186262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,7168,0.011384000380833944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,65536,0.19593600432078043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,6144,0.014243554737832812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,8192,0.015778667396969266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,768,0.013373333546850415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,1536,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,65536,0.059474666913350425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,8192,0.014269333746698169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,3072,0.006559111177921295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,1024,0.048024889495637685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,64,0.00868088917599784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,4096,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,12288,0.019863999552196927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,6144,0.012653333445390066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,4096,0.030441777573691473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,12288,0.12434755431281196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,3584,0.008088889221350351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,256,0.0032604444358083936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,128,0.0158915552828047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,256,0.005643555687533484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,2560,0.00703288863102595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,5120,0.01069777790043089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,512,0.015227556228637695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,3072,0.007153777612580194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,2560,0.07760266462961833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,16384,0.0217902229891883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,3072,0.01753422286775377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,10240,0.012338666452301873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,2048,0.006305777778228124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,2560,0.00720000018676122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,32,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,16384,0.10979466968112522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,1536,0.7449386914571127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,1024,0.008725333544943068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,65536,0.02403733299838172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,2560,0.03901155458556281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,16384,0.01499644418557485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,10240,0.026717333330048457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,1024,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,16384,0.02032800018787384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,12288,0.019674667053752475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,3072,0.030495110485288832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,512,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,64,0.0027919999427265594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,1024,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,8192,0.023629332582155865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,1536,0.01443911095460256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,3072,0.007637333538797166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,2048,0.0058044443527857465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,16384,0.06747111346986559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,16384,0.025027554896142747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,3072,0.015554666519165039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,3584,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,512,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,5120,0.08992266654968262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,1536,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,10240,0.029088000456492107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,12288,0.04407555527157254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,4096,0.014298665854665967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,768,0.016525333126386006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,128,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,65536,0.06860444280836317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,6144,0.019927110936906602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,8192,0.01962311069170634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,16384,0.020207999481095206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,2560,0.046463999483320445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,256,0.0052435555391841466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,12288,0.025303999582926433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,1536,0.010425777898894416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,768,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,3584,0.02693422304259406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,4096,0.008262222011884054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,512,0.014153778553009033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,12288,0.5115377638075087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,16384,0.10687288973066543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,2048,0.009690666364298927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,128,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,4096,0.02538577715555827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,12288,0.033012443118625216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,5120,0.0525182220670912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,1536,0.008395555946562026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,16384,0.04962044292026096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,65536,0.08484088712268406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,12288,0.008175111479229396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,7168,0.028565333949195013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,4096,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,128,0.011747555600272285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,3584,0.014341332846217685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,12288,0.02604088849491543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,16384,0.06274488899442884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,10240,0.015472888946533203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,65536,0.05011111166742113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,3072,0.00720177756415473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,12288,0.04233244392606947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,1024,0.004864000197913912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,4096,0.009369778136412302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,2048,0.0064151109092765385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,7168,0.022488888767030504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,1024,0.016840888394249808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,2560,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,12288,0.017415111263593037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,16384,0.04496444596184624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,6144,0.01962577799956004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,16384,0.12520266903771296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,65536,0.06448088751898871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,10240,0.01239377756913503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,256,0.019712888532214694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,3584,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,7168,0.010324444207880232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,4096,0.010173333187898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,16384,0.020005333754751418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,10240,0.028218666712443035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,512,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,4096,0.021848888860808477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,5120,0.018037334084510803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,2048,0.006683555742104848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,8192,0.008365333080291748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,2048,0.0141004439857271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,8192,0.010718222293588849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,256,0.005384888913896348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,6144,0.013325333595275879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,7168,0.011357333097192975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,3584,0.009993777506881291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,128,0.01648000048266517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,768,0.00869422240389718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,3072,0.012473777764373355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,128,0.00534400012758043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,1024,0.010244444840484196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,65536,1.9264240264892578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,256,0.02328799996111128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,3584,0.03593688872125413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,1024,0.00740444411834081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,2048,0.01926399932967292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,1536,0.024068444967269897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,1024,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,1024,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,3072,0.03233955634964837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,5120,0.01555022266176012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,6144,0.03554755449295044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,2048,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,3072,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,6144,0.04179733329349094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,5120,0.10416444142659505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,7168,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,7168,0.07540977663464017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,8192,0.07836977640787761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,65536,0.16697866386837432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,7168,0.020250666472646926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,1536,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,5120,0.008289777570300633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,7168,0.019845333364274766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,10240,0.0881991121504042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,7168,0.05715199973848131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,512,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,768,0.06274044513702393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,4096,0.01477777792347802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,7168,0.08640888664457534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,7168,0.03477422065205044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,128,0.016532444291644625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,16384,0.02515999972820282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,2048,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,3584,0.016508445143699646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,1536,0.009033777647548253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,7168,0.05893422497643364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,7168,0.07662844657897949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,1024,0.0041893331540955436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,6144,0.009325332939624786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,5120,0.011062222222487131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,1536,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,2560,0.014867555763986377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,6144,0.023414222730530634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,128,0.014897776974572076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,7168,0.02267733381854163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,1536,0.01388533330625958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,512,0.003539555602603488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,5120,0.1820088889863756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,128,0.02053511142730713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,3584,0.11252444320254856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,512,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,12288,0.047726223866144814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,3072,0.01585155559910668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,1024,0.016324443949593436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,3584,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,10240,0.3406702147589789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,256,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,128,0.009015111459626092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,10240,0.046670221620135836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,3584,0.034182220697402954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,256,0.01755911111831665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,3584,0.0296106669637892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,2560,0.016858667135238647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,12288,0.019202666150199044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,3584,0.01310488912794325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,3584,0.02337155573897892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,12288,0.05628089110056559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,256,0.01665422154797448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,768,0.005548444473081165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,12288,0.040772444672054715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,6144,0.00965244405799442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,10240,0.025695110360781353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,3072,0.011695110963450538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,7168,0.010775110787815519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,10240,0.012314666476514606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,128,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,3584,0.020079111059506733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,8192,0.091921779844496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,64,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,16384,1.8933662838406031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,1536,0.01649599936273363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,65536,0.24498044119940865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,2560,0.018565333551830716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,10240,0.02328266700108846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,12288,0.09241333272722031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,6144,0.07618577612770928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,768,0.01739644507567088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,65536,0.07581777705086602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,10240,0.027425777581002977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,16384,0.01759022143152025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,1024,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,2048,0.021292444732454088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,6144,0.00961866643693712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,65536,0.05914400021235148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,3584,0.024668445189793903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,3072,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,6144,0.013388444152143268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,7168,0.012773333324326409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,64,0.010016889207892949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,128,0.04854399959246317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,7168,0.017044444878896076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,1024,0.017820444371965196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,32,0.004925333377387789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,6144,0.19460977448357475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,768,0.01586844523747762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,65536,0.28987110985649955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,2048,0.023354666100607976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,2048,0.04861155483457777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,1536,0.0063440000845326324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,32,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,768,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,5120,0.00834222220712238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,4096,0.013906665974193148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,128,0.0034862222770849862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,8192,0.024039111203617517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,65536,0.07168266508314344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,3072,0.013821333646774292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,12288,0.03636533353063796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,16384,0.03836799992455377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,5120,0.015146666102939181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,16384,0.05579644441604614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,3072,0.007312888900438945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,10240,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,2048,0.005647999958859549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,5120,0.013814222481515674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,3072,0.01757066614098019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,1024,0.013748444616794586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,65536,0.017853332890404593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,65536,0.23377066188388398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,3072,0.04062044289377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,8192,0.03188177943229675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,6144,0.023394667439990576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,65536,0.06315288941065471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,3072,0.12048799461788601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,5120,0.02369955513212416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,12288,0.03523555397987366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,768,0.017316444052590262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,512,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,8192,0.01939022209909227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,5120,0.016693333784739178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,1024,0.01688088807794783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,768,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,10240,0.01461155547036065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,10240,0.05951999955707126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,1024,0.005688000056478713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,4096,0.025609778033362493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,12288,0.04520444406403435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,3584,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,8192,0.017312000195185345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,2560,0.007232888705200619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,7168,0.010245333115259806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,128,0.0035173334181308746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,3072,0.01365244471364551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,65536,0.059770663579305015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,7168,0.020598222812016804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,768,0.0069724445541699724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,7168,0.0313937763373057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,16384,0.12363378206888835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,512,0.003536000019974179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,1536,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,12288,0.026195555925369263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,2048,0.01962488889694214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,6144,0.02292711039384206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,3072,0.23566754659016928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,768,0.016527111331621807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,3584,0.3513022263844808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,65536,0.11209511756896973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,65536,0.0837964415550232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,128,0.02313599983851115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,128,0.014832889040311178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,4096,0.0217902229891883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,5120,0.03154133425818549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,4096,0.011569778124491373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,1024,0.006451555424266391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,8192,0.014890667464998035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,32,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,1536,0.013847110999955071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,3584,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,3072,0.015862221519152325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,12288,0.01718577742576599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,7168,0.01278133359220293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,10240,0.032383110788133405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,256,0.043381333351135254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,1536,0.021498666869269475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,12288,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,128,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,4096,0.03208355440033807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,32,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,256,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,6144,0.0299928883711497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,4096,0.0219217770629459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,65536,0.06402755445904203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,10240,0.03548533386654324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,4096,0.01819022165404426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,5120,0.009016888837019602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,512,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,4096,0.03258488906754388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,256,0.0057795556883017225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,1024,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,32,0.007252444823582967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,5120,0.048397332429885864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,16384,0.04899200134807163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,6144,0.0537066658337911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,128,0.01276266657643848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,65536,0.23979287677341035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,12288,0.021592888567182753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,256,0.0031884445084465873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,2560,0.03946933481428359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,1024,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,65536,0.08593422174453735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,6144,0.3200604385799832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,128,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,65536,0.07986133628421359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,1536,0.00453599997692638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,12288,0.01958044370015462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,128,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,3072,0.16320978270636663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,65536,0.07319555679957072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,3584,0.021158221695158217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,3072,0.006955555743641323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,1024,0.014135110709402295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,128,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,3072,0.00907466643386417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,1536,0.01237777786122428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,1024,0.004592888885074192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,16384,0.023415999280081853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,1536,0.01828799976242913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,8192,0.010397333237859938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,6144,0.0368755559126536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,16384,0.04741955465740628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,3584,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,12288,0.20206756061977812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,7168,0.10693955421447754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,768,0.016418667303191293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,1024,0.018574222922325134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,6144,0.00834222220712238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,512,0.3707679907480876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,6144,0.009355555805895064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,6144,0.04134133458137512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,512,0.009688888986905416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,2560,0.024944000773959692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,1536,0.012629333469602795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,3584,0.03674133287535773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,512,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,128,0.0028604444944196274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,7168,0.07233866718080309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,768,0.013737777868906656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,7168,0.03647644321123759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,16384,0.1437573300467597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,12288,0.02296977738539378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,8192,0.03905244337187873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,128,0.07038488652971056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,8192,0.017839110559887357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,3584,0.07244444555706449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,512,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,128,0.003112888791494899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,1536,0.005985777825117111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,4096,0.009279111193286048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,2560,0.007271111011505127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,10240,0.016174222032229107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,8192,0.023400889502631292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,5120,0.019176888797018263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,3584,0.040159109565946795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,2560,0.07975822024875216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,8192,0.01645777788427141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,1024,0.012109333442317115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,768,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,6144,0.021651556094487507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,1024,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,1024,0.025403555896547105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,12288,0.018608889646000333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,3072,0.03659910957018534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,3584,0.040122667948404946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,6144,0.018611555298169453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,65536,0.07347466548283894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,16384,0.01720266706413693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,12288,0.015726221932305228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,4096,0.009338666995366415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,16384,0.01650044487582313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,2560,0.025636444489161175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,8192,0.03818755679660373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,512,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,6144,0.02084622283776601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,1024,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,3072,0.007592889169851939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,64,0.047285334931479565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,12288,0.02125511070092519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,3072,0.026113778352737427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,4096,0.01754755609565311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,16384,0.02453688946035173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,768,0.06642844279607137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,3072,0.017110221915774874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,12288,0.025452444950739544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,8192,0.010771555205186209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,3584,0.008105778031879002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,16384,0.023922666907310486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,2560,0.007617777420414819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,3584,0.008629333641793992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,7168,0.12082844310336643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,2048,0.07397511270311144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,768,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,32,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,1536,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,16384,0.032590223683251276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,12288,0.038952890369627215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,1536,0.007490666376219855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,1024,0.009308444129096137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,7168,0.019256888164414298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,16384,0.025386666258176167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,5120,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,256,0.0032035555276605818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,3072,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,64,0.004594666676388847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,10240,0.1952364444732666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,12288,0.04877777894337972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,128,0.015494222442309061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,4096,0.02275288932853275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,32,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,5120,0.11254044373830159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,256,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,65536,0.21067733234829375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,8192,0.015218666858143277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,10240,0.017218665944205392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,768,0.008347555167145198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,256,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,768,0.00793599999613232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,12288,0.18498044543796113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,1536,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,256,0.003185777821474605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,10240,0.2714933289421929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,4096,0.19346933894687227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,12288,0.028539554940329656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,8192,0.012457778056462606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,6144,0.016923555069499545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,768,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,3072,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,8192,0.02200177808602651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,16384,0.03123377760251363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,768,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,3584,0.016125332978036668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,1024,0.004855999930037392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,65536,0.3524773385789659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,4096,0.032355556885401406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,2048,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,7168,0.02334311107794444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,768,0.00452533322903845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,128,0.00313688897424274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,256,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,10240,0.032953778902689614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,65536,0.08330221970876057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,256,0.0038542221817705366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,1024,0.0613280004925198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,7168,0.019908444748984445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,2048,0.0069057775868309875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,256,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,1536,0.013782222237851886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,65536,0.1655671066708035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,1024,0.007871111234029135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,8192,0.03350399931271871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,10240,0.02409066590997908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,128,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,32,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,2048,0.008642666869693333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,768,0.005032888717121548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,512,0.0148053334818946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,128,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,128,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,7168,0.02862311071819729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,6144,0.24281867345174155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,5120,0.048698667022917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,2048,0.00870400004916721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,8192,0.016171556380059984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,2048,0.013256000147925483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,4096,0.015837333268589444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,10240,0.02681066592534383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,768,0.0038746665749284956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,5120,0.007594666547245449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,5120,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,1536,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,1536,0.012097777591811286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,1536,0.0075048887067370946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,6144,0.017924444542990792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,3072,0.015604444675975375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,1024,0.04399111204677158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,8192,0.01683999929163191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,2560,0.023037332627508376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,256,0.003456888927353753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,1536,0.008626666333940294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,65536,0.31630222002665204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,1536,0.005146666533417172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,16384,0.05593066745334201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,10240,0.017027555240525138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,1024,0.0063279999627007385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,768,0.009000000026490953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,12288,0.03252089023590088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,128,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,4096,0.5914799902174208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,65536,0.1307635572221544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,2048,0.009713778065310584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,1536,0.010829333629873065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,768,0.04965600040223864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,1536,0.026148445076412622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,512,0.01585244470172458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,1024,0.01237511138121287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,1536,0.021987555755509272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,3072,0.05826489130655924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,256,0.04734666811095344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,1536,0.03473511007097032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,6144,0.028353777196672227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,65536,0.015378667248619927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,512,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,2048,0.032109333409203425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,65536,0.13673688305748835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,4096,0.02783999840418498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,65536,0.015086222026083203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,3584,0.017456889152526855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,64,0.004517333375083076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,512,0.01513066722287072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,7168,0.019690665933820937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,128,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,12288,0.02975289026896159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,1536,0.00741244438621733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,2560,0.007373332977294922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,10240,0.022292445103327434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,2048,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,256,0.016547555724779766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,1024,0.014799111419253878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,768,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,2560,0.010192888478438059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,32,0.011014222270912595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,2560,0.01680799987581041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,3584,0.061928888161977134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,7168,0.031834665271970965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,128,0.01551999979548984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,16384,0.032572444942262434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,1024,0.007160888777838812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,12288,0.658342202504476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,1536,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,12288,0.01819022165404426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,8192,0.04794489012824165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,768,0.01651822196112739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,65536,0.06291733185450236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,512,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,1536,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,7168,0.03017688790957133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,16384,0.1961644490559896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,512,0.0052791109515561005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,7168,0.35137067900763613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,64,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,12288,0.03600800037384033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,256,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,256,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,5120,0.015944888194402058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,5120,0.021176000436147053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,5120,0.048997332652409874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,16384,0.05097600155406528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,10240,0.017456000049908955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,16384,0.03146933184729682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,3584,0.039332442813449435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,64,0.007603555917739868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,1536,0.01108888867828581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,5120,0.010064000056849586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,16384,0.044028444422615894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,4096,0.01149511088927587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,16384,0.08269511328803168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,128,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,1024,0.01310755560795466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,256,0.003003555453485913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,3072,0.012472888661755456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,768,0.015156444576051501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,16384,0.08642044332292344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,5120,0.019277332557572257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,65536,0.06235288911395603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,10240,0.10988622241550022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,768,0.008952000074916417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,64,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,256,0.016885333591037326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,5120,0.01835555500454373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,65536,0.013810666898886362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,5120,0.10407555765575832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,1536,0.00702133360836241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,128,0.015431111057599386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,65536,0.10257066620720758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,10240,0.02793688906563653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,512,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,64,0.005744888964626525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,256,0.003010666618744532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,2560,0.04302666584650675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,8192,0.013996443814701505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,5120,0.03016177813212077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,1536,0.044256889157825045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,3072,0.021827555365032617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,6144,0.012804444465372296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,5120,0.01620444489849938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,65536,0.3024097813500298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,1024,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,512,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,8192,0.08553510904312134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,3584,0.013025778035322825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,2048,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,128,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,512,0.01276000009642707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,256,0.0028373334142896864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,768,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,6144,0.06779111093944974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,1536,0.020975111259354487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,32,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,512,0.014836443795098199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,1024,0.00407911092042923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,8192,0.05252266592449612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,6144,0.009485333330101436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,3072,0.00998399986161126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,8192,0.013460444079505073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,2048,0.011525332927703857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,768,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,2048,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,64,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,12288,0.04882488979233635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,5120,0.009076444639099969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,12288,0.019301333361201815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,12288,0.0240639994541804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,512,0.02329866588115692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,5120,0.014340443743599786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,768,0.004524444540341695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,10240,0.0165955556763543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,512,0.006014222072230444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,7168,0.011607111328177981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,512,0.003253333270549774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,5120,0.041618665059407554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,256,0.029289778735902574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,256,0.009101333717505137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,8192,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,3072,0.007288888924651676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,2560,0.2606595622168647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,64,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,10240,0.2256631056467692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,512,0.003687111039956411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,4096,0.01686755485004849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,65536,0.06584799951977201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,5120,0.008676444490750631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,512,0.005850666513045629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,256,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,6144,0.1145626703898112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,8192,0.010761777559916178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,65536,0.17657244205474854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,6144,0.03268622358640035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,12288,0.013351110948456658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,2048,0.04153688748677572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,65536,0.018087999688254464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,12288,0.01310400002532535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,10240,0.04800533254941305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,3584,0.011771555576059552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,768,0.0052177777720822235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,10240,0.017501332693629794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,512,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,2560,0.015512888630231222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,2048,0.006551110910044775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,64,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,10240,0.015961777832772996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,3584,0.028714666763941448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,2048,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,6144,0.04284888837072584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,10240,0.02130399975511763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,512,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,3072,0.048751112487581044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,5120,0.005964444329341252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,3072,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,1024,0.008287111090289222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,12288,0.10763377613491482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,8192,0.01107022249036365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,1536,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,2048,0.006994666324721442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,6144,0.018351111147138808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,7168,0.023388443721665278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,3072,0.016128889388508268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,65536,0.036906666225857206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,4096,0.09500088956620957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,3072,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,5120,0.02127733329931895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,32,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,7168,0.01477777792347802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,768,0.01756177842617035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,128,0.0214720004134708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,128,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,512,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,6144,0.00903911143541336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,2560,0.054400000307295054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,7168,0.016506666938463848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,6144,0.010334221853150262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,1536,0.006548444430033366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,6144,0.013452444639470843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,2560,0.010418666733635796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,3584,0.013369777964221107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,4096,0.011359111302428775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,1024,0.014448000325096978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,16384,0.03014844324853685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,4096,0.023061333431137934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,64,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,8192,0.022073777185546026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,8192,0.026106667187478807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,4096,0.028892444239722356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,6144,0.013564444250530668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,32,0.0069724445541699724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,3584,0.05534311135609945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,4096,0.03011288907792833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,8192,0.24262934260898164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,2560,0.010808889236715106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,2560,0.009673777553770278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,6144,0.12632532914479574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,10240,0.014505777094099255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,6144,0.014122666584120857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,512,0.01316088851955202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,8192,0.05784622165891859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,768,0.0074577778577804565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,32,0.0032995556377702286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,5120,0.02257777750492096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,16384,0.01759199963675605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,64,0.0038151111867692736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,7168,0.032038221756617226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,7168,0.028020444843504164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,5120,0.01583377851380242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,12288,0.1782595581478543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,65536,0.36763821707831484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,7168,0.02162933349609375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,1536,0.030393779277801514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,65536,0.08406400018268162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,3072,0.021087111698256597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,6144,0.019235556324323017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,3584,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,128,0.014177777700954013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,3072,0.007262222468852997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,1536,0.054288887315326266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,3072,0.038986666334999934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,4096,0.03428533342149522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,3584,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,512,0.005649777750174205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,4096,0.009350222018029954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,10240,0.011753777662913004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,8192,0.01575644479857551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,5120,0.016639111770523917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,3072,0.03253600001335144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,4096,0.008961777720186446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,16384,0.047395557165145874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,3584,0.017637333936161466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,6144,0.17995022402869332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,2048,0.016153777639071148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,1024,0.0063235556913746735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,256,0.011021333436171213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,65536,0.016694222887357075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,1024,0.0116462219092581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,5120,0.01309866706530253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,10240,0.08061777883105807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,768,0.01682133310370975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,3072,0.011439111497667102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,768,0.007975999679830339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,2048,0.03597955571280585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,7168,0.01611022154490153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,768,0.016275554895401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,2048,0.0199297782447603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,128,0.0038151111867692736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,16384,0.03436622354719374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,4096,0.052870220608181424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,2048,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,10240,0.20589066876305473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,10240,0.024163555767801072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,768,0.00490133340160052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,65536,0.05614044268925985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,2048,0.012734221915404001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,6144,0.011759999725553723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,5120,0.10795644256803726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,6144,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,2560,0.018933332628673978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,2560,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,65536,0.1512462165620592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,3584,0.013541333377361298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,6144,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,64,0.00305244450767835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,512,0.01959111127588484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,8192,0.025628444221284654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,256,0.007602666815121968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,3584,0.007680888805124495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,16384,0.029198222690158423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,65536,0.08025244209501478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,32,0.006259555617968242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,768,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,2560,0.011457777685589261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,10240,0.029042667812771265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,512,0.012895111408498553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,7168,0.02405511173937056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,2560,0.00794666674402025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,1024,0.02093155516518487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,64,0.0028728888266616394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,7168,0.06018400192260742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,768,0.003625777860482534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,768,0.003855111284388436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,7168,0.26480356852213544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,10240,0.027221333649423387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,1536,0.013463111387358772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,65536,0.628861321343316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,6144,0.009027555584907532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,128,0.018214222457673814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,1536,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,10240,0.0058559998869895935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,12288,0.0306693348619673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,768,0.34286755985683864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,12288,0.010986666712496014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,7168,0.6246799892849392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,3584,0.01651822196112739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,7168,0.0311173333062066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,6144,0.28971555497911244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,1536,0.010684444672531553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,5120,0.011067555182509951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,128,0.024315555890401203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,768,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,128,0.015134221977657743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,16384,0.021475555168257818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,16384,0.02529866662290361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,12288,0.016826666063732572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,2048,0.01666844387849172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,768,0.0035075553589397003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,2048,0.13559288448757595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,1536,0.006228444476922353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,2560,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,2560,0.0069013333155049225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,12288,0.026399999856948853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,128,0.01477599971824222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,12288,0.028613332245084975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,12288,0.022554667459593877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,8192,0.0162462227874332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,7168,0.05946933560901218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,1536,0.0063368889192740125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,65536,0.29765955607096356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,16384,0.01586933268441094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,32,0.007965332931942409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,512,0.005234666582610872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,7168,0.039953778187433876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,5120,0.03526844580968221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,768,0.007596444752481248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,10240,0.017131555411550734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,8192,0.019574221637513902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,3584,0.007908444437715743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,8192,0.35097334120008683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,128,0.007642666498819987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,64,0.005979555762476391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,65536,0.36932354503207737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,256,0.03927377859751383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,1536,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,512,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,2560,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,10240,0.10588000218073527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,6144,0.03204266561402215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,2560,0.014507555299335055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,2560,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,10240,0.16565244727664524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,4096,0.06528888808356391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,1024,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,1024,0.01645599967903561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,2560,0.03171111146608988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,64,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,1536,0.00685777763525645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,12288,0.039289779133266874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,2048,0.008023999631404877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,6144,0.02682222260369195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,32,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,32,0.002699555622206794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,4096,0.02860622273551093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,2048,0.07515022489759657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,512,0.003495111233658261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,64,0.0042133331298828125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,7168,0.011765333513418833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,6144,0.010704889065689512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,7168,0.009501333038012186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,12288,0.14947733614179823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,2048,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,65536,0.0785964462492201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,10240,0.0782817800839742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,32,0.003113777687152227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,4096,0.00831822223133511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,12288,0.01993600030740102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,16384,0.01681511104106903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,512,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,32,0.007860444486141205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,512,0.01276000009642707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,10240,0.012078222301271228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,32,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,3072,0.022554667459593877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,768,0.015188443991872998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,7168,0.09921333524915908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,1536,0.006385777973466449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,768,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,3072,0.020297777321603563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,512,0.003576888806290097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,6144,0.026334222820070054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,65536,0.04945866598023308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,512,0.012777778009573618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,8192,0.034861332840389676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,5120,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,8192,0.010739555789364709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,16384,0.015802666544914246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,12288,0.1818151076634725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,65536,0.02549422283967336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,65536,0.17630933390723336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,65536,0.17334489027659097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,2048,0.007258666886223688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,256,0.03247200118170844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,768,0.014787554740905762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,768,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,16384,0.023883556326230366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,128,0.014765333798196582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,10240,0.03804355528619554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,12288,0.022608000371191237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,32,0.00553955551650789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,7168,0.018920888503392536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,256,0.024911999702453613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,3584,0.007688889073001013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,768,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,2560,0.02420622275935279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,4096,0.06111555629306369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,12288,0.037296000454160906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,4096,0.02808444367514716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,2560,0.008791999684439765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,128,0.04281333420011732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,1536,0.004753777964247597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,16384,0.04504533277617561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,7168,0.45978842841254336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,768,0.008032889001899296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,2560,0.01829688913292355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,4096,0.017823111679818895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,2048,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,16384,0.04261066516240438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,3584,0.013153778182135688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,7168,0.009702222214804756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,1536,0.02372444503837162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,65536,0.08257689078648885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,1024,0.0038764443662431505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,3584,0.013464888764752282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,7168,0.07654755645328097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,1536,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,5120,0.01672266589270698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,2048,0.012200888660218982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,512,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,2048,0.011032889286677042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,3072,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,6144,0.021402666966120403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,1536,0.014058666096793281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,512,0.003564444267087512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,12288,0.3184915648566352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,768,0.003546666767862108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,512,0.005108444641033809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,128,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,128,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,3584,0.12365244494544135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,2048,0.01036000003417333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,3584,0.04063466522428725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,1024,0.020488889680968392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,7168,0.05434222353829277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,512,0.0035582222044467926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,2048,0.005767111149099138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,65536,0.12420533763037787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,32,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,2560,0.012470222181744046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,2048,0.08143911096784803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,16384,0.04014933440420363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,256,0.008347555167145198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,5120,0.021716444028748408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,65536,0.13183200359344482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,768,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,4096,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,5120,0.02712000078625149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,8192,0.00700711127784517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,12288,0.011422221859296164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,3072,0.007781333393520779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,1536,0.007672000262472365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,2048,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,7168,0.020978666014141504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,7168,0.019897777173254225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,256,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,1024,0.004306666553020477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,8192,0.05673422416051229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,3072,0.015542222393883599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,7168,0.010288889209429422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,12288,0.4944471253289117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,256,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,10240,0.025105777713987563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,128,0.0035626664757728577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,2048,0.028350220786200628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,12288,0.010844444235165914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,6144,0.02340000040001339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,10240,0.009094222552246517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,3584,0.0362062222427792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,256,0.003624888757864634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,768,0.0064586665895250105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,1024,0.0053075556125905775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,3584,0.010763555765151978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,8192,0.014150222142537435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,128,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,4096,0.007873777714040544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,8192,0.011201777391963534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,1536,0.040431999497943454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,4096,0.013235555754767524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,16384,0.012979555461141797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,5120,0.022473777333895367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,16384,0.022279999322361414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,1536,0.014588443769348992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,16384,0.04853066802024841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,32,0.007897777689827813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,256,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,1024,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,2048,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,768,0.005983110931184556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,10240,0.019875556230545044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,1024,0.016591999265882704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,65536,0.27047822210523814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,768,0.003926222109132343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,3072,0.012647999657524956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,64,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,6144,0.02553422252337138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,3584,0.00870311094654931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,1536,0.008288000192907121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,12288,0.025028443998760645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,7168,0.056114666991763644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,3072,0.027035555905765955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,256,0.022281777527597215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,2560,0.007672000262472365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,2048,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,256,0.02344977855682373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,8192,0.0403422216574351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,256,0.014813333749771118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,512,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,10240,0.02531822191344367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,768,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,6144,0.19267822636498344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,2048,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,2048,0.07770133018493652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,768,0.007294221884674496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,1024,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,8192,0.010436444646782346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,7168,0.15168889363606772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,1536,0.018600889378123812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,7168,0.020285333196322124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,2048,0.009692444569534725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,512,0.0041697778635554844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,64,0.0034959999223550162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,5120,0.058168000645107694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,1024,0.02628533376587762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,65536,0.07652088668611315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,10240,0.029976000388463337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,1536,0.00741155528359943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,7168,1.2810711330837674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,32,0.020271110865804885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,3584,0.007630222373538547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,65536,18.19567701551649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,10240,0.03292444348335266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,65536,0.09858044650819565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,32,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,8192,0.024079110887315538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,2560,0.01112711098459032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,2048,0.014135999812020196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,128,0.02160355614291297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,10240,0.05682933330535889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,6144,0.025342222717073228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,256,0.007369777394665613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,32,0.0029022221763928733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,10240,0.2599155637953016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,2048,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,768,0.017113778326246474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,2048,0.006974221931563483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,16384,0.022310222188631695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,10240,0.023588443795839947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,512,0.016879111528396606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,1024,0.011751111182901593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,8192,0.24356532096862793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,16384,0.09049866596857707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,65536,0.060518225034077965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,1024,0.037618666887283325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,32,0.0028159999185138275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,3584,0.012905778156386482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,32,0.0028657778683635923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,5120,0.030568000343110826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,8192,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,3584,0.007178666690985362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,4096,0.008372444245550368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,1024,0.0144177774588267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,65536,0.06980799966388278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,4096,0.01568711135122511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,3584,0.0076453329788313965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,10240,0.05480622251828512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,5120,0.1654097768995497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,8192,0.010400888820489248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,768,0.00620977787507905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,128,0.011900444825490316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,2560,0.007897777689827813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,768,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,7168,0.011576889289749993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,16384,0.025815111067559984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,12288,0.03045511245727539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,6144,0.0898959967825148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,10240,0.03543466660711501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,128,0.0029031110720502008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,256,0.0031786666562159858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,256,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,16384,0.042529778348075024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,64,0.007685333490371704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,1536,0.016385777129067313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,2048,0.02752444479200575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,2560,0.0215013325214386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,8192,0.011003555523024665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,7168,0.0454728901386261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,12288,0.01309866706530253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,4096,0.007782222496138678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,5120,0.008323555191357931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,128,0.04284888837072584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,1024,0.0058293334311909145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,1536,0.09126310878329807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,7168,0.012090666426552666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,2048,0.015804444750150044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,512,0.01816177699300978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,64,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,10240,0.03612622287538316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,512,0.0035297779573334586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,6144,0.018263111511866253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,1024,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,128,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,512,0.0052595556610160405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,8192,0.029665778080622356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,1536,0.028776887390348647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,3072,0.007791111038790808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,3584,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,8192,0.04525422387652927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,6144,0.03934311204486423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,16384,0.013832000394662222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,768,0.05032711227734884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,768,0.018959111637539335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,32,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,7168,0.012529777983824411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,256,0.0028968888024489083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,512,0.006697777658700943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,256,0.01720977822939555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,65536,0.08288177516725329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,5120,0.025116443634033203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,5120,0.00834222220712238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,6144,0.009309333231714036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,512,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,3072,0.015996444556448195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,1536,0.004868444469239977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,7168,1.9686595069037542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,12288,0.16036089261372885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,256,0.005677333308590784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,512,0.005072000126043956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,768,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,10240,0.01588000026014116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,3584,0.02785244584083557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,512,0.003543111185232798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,64,0.003128888913326793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,4096,0.010447111394670276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,1536,0.006565333240562015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,768,0.004861333303981357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,12288,0.05665510892868042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,128,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,6144,0.01748533381356133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,12288,0.037547555234697126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,1024,0.0064399999876817065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,512,0.010412444670995077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,2048,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,5120,0.009903111391597325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,16384,0.015792888071801927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,2560,0.11222755908966064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,16384,0.0628239976035224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,12288,0.10763733254538642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,12288,0.01863822175396813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,128,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,6144,0.027383110589451257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,3584,0.007015110717879401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,6144,0.01992266707950168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,64,0.0025235555238193935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,7168,0.020377778344684176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,5120,0.008600888980759515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,768,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,2048,0.036004443963368736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,65536,0.08422133657667373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,7168,0.018219555417696636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,2048,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,6144,0.0090515555606948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,10240,0.022646221849653456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,5120,0.00961688905954361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,1536,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,12288,0.03269066744380527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,3584,0.009319110876984065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,2048,0.013463111387358772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,512,0.01478666729397244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,256,0.006416888700591193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,4096,0.008398222426573435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,2048,0.012450666891203986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,65536,0.02696888811058468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,6144,0.039192001024881996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,3584,1.752462175157335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,768,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,8192,0.04651022288534376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,512,0.006160000132189856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,1024,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,3072,0.01698755555682712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,4096,0.009455111291673448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,12288,0.9148160086737739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,512,0.01646933290693495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,256,0.02439111140039232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,5120,0.026202667090627883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,1024,0.015097777048746744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,3072,0.007745777567227681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,1536,0.10339111089706421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,1024,0.013767999907334646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,3072,0.014484445254007975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,32,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,128,0.00527022240890397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,65536,0.06591466400358412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,2560,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,6144,0.015578667322794596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,1536,0.006226666685607698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,7168,0.009025777379671732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,1024,0.017893332574102614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,7168,0.012315555579132505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,64,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,5120,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,10240,0.023846222294701472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,8192,0.17889155281914604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,6144,0.020238222347365487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,1024,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,65536,0.046463999483320445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,2560,0.012208888928095499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,65536,0.11036088731553818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,64,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,3072,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,256,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,1024,0.0181057784292433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,10240,0.01973511113060845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,256,0.006425777657164468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,768,0.007999111380841997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,768,0.004176889028814104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,1024,0.018090666996108163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,5120,0.008299555215570662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,7168,0.013438222308953604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,128,0.015493333339691162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,3584,0.009335111412737105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,64,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,3584,0.025935111774338618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,1536,0.012438221938080259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,64,0.0029084444459941653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,7168,0.03258488906754388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,768,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,12288,0.12329867151048447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,4096,0.028696000576019287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,512,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,4096,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,512,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,64,0.005761777775155173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,768,0.016807110773192514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,1024,0.07207644648022123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,3584,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,512,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,32,0.009716444545321995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,768,0.01994755533006456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,2048,0.16131733523474798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,4096,0.015480889214409722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,512,0.022351110974947613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,5120,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,7168,0.014278221461508008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,7168,0.014153778553009033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,65536,9.290771484375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,10240,0.03211911188231574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,16384,0.032227555910746254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,10240,0.21008977625105116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,12288,0.01344622257683012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,512,0.012789333032237159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,64,0.010424888796276517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,3584,0.007974222302436829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,5120,0.009268444445398118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,8192,0.03823022378815545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,8192,0.4048515425788031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,512,0.008991111483838823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,2048,0.010120888551076254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,512,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,16384,0.034019556310441755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,2048,0.014326221413082547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,10240,0.026125333375400964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,4096,0.022774222824308608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,512,0.006425777657164468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,6144,0.01758577757411533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,256,0.0031600000543726813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,32,0.0041866666740841335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,5120,0.006212444355090459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,12288,0.01310400002532535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,32,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,512,0.010471111370457543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,1024,0.01514488955338796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,32,0.003113777687152227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,10240,0.007788444558779399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,16384,0.22628710005018446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,4096,0.010429333481523724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,4096,0.02345422241422865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,65536,0.07234844234254625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,768,0.00774133370982276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,3584,0.018654222289721172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,6144,0.018228444788191054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,12288,0.013110222087966071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,5120,0.03462222218513489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,10240,0.02106844385464986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,2560,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,1024,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,256,0.005211555709441503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,12288,0.021935999393463135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,3072,0.03267466690805223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,256,0.005879999862776862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,3584,0.023430221610599097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,2048,0.01312888910373052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,128,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,7168,0.009384888741705153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,768,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,2560,0.005673777725961473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,512,0.003633777714437909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,1024,0.0039057777159743835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,7168,0.01312888910373052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,65536,0.27616532643636066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,3072,0.11692710717519124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,7168,0.014630221658282809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,4096,0.0471093323495653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,7168,0.016099555624855887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,7168,0.020262221495310467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,12288,0.031142221556769475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,10240,0.02534399926662445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,8192,0.029483556747436523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,768,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,3072,0.0073315559162033936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,3584,0.016872000363137987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,2048,0.01626488897535536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,4096,0.03616444600952996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,8192,0.028742220666673448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,5120,0.018544000056054857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,1024,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,16384,0.023630221684773762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,1024,0.00794488853878445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,512,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,2048,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,5120,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,8192,0.027235555979940627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,768,0.00675999994079272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,2560,0.03156355685657925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,3072,0.0099982221921285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,8192,0.14579910702175564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,1536,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,2560,0.08311733272340563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,4096,0.011943110989199745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,5120,0.16506400373246935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,768,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,256,0.0232640008131663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,512,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,256,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,256,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,1024,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,4096,0.2202462222841051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,10240,0.012073777616024017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,8192,0.027103111147880554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,4096,0.010031999813185798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,2560,0.010144000252087912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,4096,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,2048,0.006351111249791251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,7168,0.01351111133893331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,10240,0.027781334188249376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,10240,0.028431998358832464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,3072,0.009359999663299983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,3584,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,65536,0.06417866547902425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,1536,0.02243999971283807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,1024,0.010411555568377176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,3072,0.014771555860837301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,4096,0.02993866801261902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,6144,0.010785777535703448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,6144,0.018408889571825664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,512,0.005447111195988125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,2048,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,256,0.005435555759403441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,12288,0.2396950986650255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,1536,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,2048,0.013156444662147097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,6144,0.013917333549923368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,768,0.01664266652531094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,2560,0.007296000089910295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,512,0.016696888539526198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,7168,0.029378665818108454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,2560,0.10550577772988214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,65536,0.42900888125101727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,10240,0.017431111799346078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,2048,0.00703288863102595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,8192,0.024314666787783306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,12288,0.01352266636159685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,64,0.007019555403126611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,3072,0.007558222446176741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,1024,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,2560,0.011072888970375061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,3584,0.042632887760798134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,256,0.00318755561278926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,8192,0.033379554748535156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,16384,0.016237333416938782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,4096,0.006970666348934174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,3072,0.046224001381132335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,6144,0.07603289021386041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,1024,0.01955200069480472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,1536,0.007278222176763747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,1536,0.004918222212129169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,8192,0.02363555630048116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,3072,0.015635554989178974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,7168,0.009696000152164036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,3072,0.007601777712504069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,3072,0.010381333529949188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,10240,0.050712890095180936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,7168,0.03412977854410807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,2560,0.13743466801113552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,12288,0.05347822109858195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,10240,0.01369955556260215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,7168,0.020356444848908316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,768,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,5120,0.023032888770103455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,16384,0.06575377782185872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,12288,0.03151022063361274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,2560,0.007251555720965068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,768,0.008654221892356873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,256,0.0034826666944556763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,512,0.013054221868515015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,3584,0.009010666774378883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,128,0.023394667439990576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,10240,0.028639998700883653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,1536,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,3072,0.016114667057991028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,16384,0.02251022226280636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,12288,0.24927287631564668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,3072,0.008164444731341468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,32,0.002798222212327851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,65536,0.020587555236286588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,768,0.007293333609898885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,512,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,5120,0.017518222332000732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,3072,0.012069332930776807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,1024,0.02951822347111172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,768,0.008741333252853816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,7168,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,1536,0.014160000615649752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,5120,0.011191111471917896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,7168,0.06312533219655354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,8192,0.03054222133424547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,128,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,1024,0.06081777811050415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,32,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,256,0.00666844430896971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,1536,0.0070088886552386815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,1536,0.07661600245369805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,64,0.007836444510353936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,2048,0.023709333605236475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,7168,0.03346577617857192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,3072,0.007608888877762689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,768,0.0037342223028341928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,32,0.0038782221575578055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,10240,0.018580444984965853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,65536,0.06191555658976237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,128,0.016185777054892648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,3072,0.01518311103185018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,10240,0.010284444524182213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,3072,0.018235555953449674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,128,0.01037599974208408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,12288,0.02922044528855218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,6144,0.02994577752219306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,3584,0.04208266735076904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,768,0.006226666685607698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,5120,0.02495199938615163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,10240,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,10240,0.012443555725945367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,768,0.017879999346203275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,128,0.0028008888992998335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,5120,0.022330666581789654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,256,0.005515555540720622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,1024,0.0058542220956749385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,4096,0.010123555858929953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,16384,0.026886221435334947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,768,0.006727111008432176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,512,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,12288,0.023027555810080633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,512,0.013752000199423896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,2560,0.0971191128094991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,8192,0.010647999743620554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,2048,0.025164445241292317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,12288,0.5552177959018284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,512,0.00675022229552269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,1536,0.00812000036239624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,4096,0.10035377740859985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,2048,0.041533334387673274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,65536,0.1783146725760566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,1024,0.008009778128729926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,10240,0.04374222291840447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,1536,0.01549600064754486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,5120,0.008588444855478074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,5120,0.008656889200210571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,256,0.003160888950030009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,7168,0.013042666845851474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,12288,0.017873777283562552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,32,0.002850666642189026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,8192,0.021293333835071985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,1536,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,2048,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,3072,0.03207644489076402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,2048,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,4096,0.1018506685892741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,768,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,3072,0.013863110707865821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,64,0.006812444577614467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,768,0.0035253332720862497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,3584,0.008941333327028487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,64,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,768,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,64,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,7168,0.031210667557186548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,5120,0.008658666577604082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,10240,0.1399093336529202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,6144,0.017854221993022494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,128,0.004182222402758068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,5120,0.037893331713146634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,5120,0.009777777724795872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,65536,0.093176888095008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,12288,0.014107555150985718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,2560,0.01331733332739936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,3072,0.050845334927241005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,2048,0.014117333624098035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,128,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,512,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,256,0.01475911173555586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,16384,0.029492444462246362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,3584,0.014866666661368476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,256,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,3584,0.01367644468943278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,7168,0.021990221407678392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,128,0.01445777714252472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,6144,0.011887110769748688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,2560,0.030177778667873804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,12288,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,5120,0.01788444485929277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,1024,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,6144,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,3584,0.015171556009186639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,7168,0.02847733431392246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,128,0.013077333569526672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,3072,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,8192,0.008995555341243744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,10240,0.03345955411593119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,65536,0.20714933342403838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,3072,0.015412444869677225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,2560,0.0070222218831380205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,4096,0.015455999308162265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,256,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,512,0.012448000411192576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,3072,0.007607111500369177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,512,0.007111999723646376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,1024,0.0075937774446275495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,256,0.005558222118351195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,256,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,1024,0.024384000235133704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,1024,0.021159110797776118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,8192,0.015754666593339708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,5120,0.01761155492729611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,512,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,2560,0.01203733351495531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,128,0.0028524444335036804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,16384,0.03629600008328756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,4096,0.01070577816830741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,256,0.0035422220826148987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,3584,0.010779555473062726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,1024,0.015807999504937064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,768,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,3072,0.44492265913221574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,16384,0.0365955564710829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,65536,0.10636444224251641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,65536,0.09316533141665989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,128,0.024681778417693243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,128,0.002814222127199173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,6144,0.05440800057517158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,5120,0.03400266501638625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,16384,0.015853333804342482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,65536,0.07704444726308186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,10240,0.025889777474933203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,6144,0.030731555488374498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,512,0.021532444490326777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,1536,0.013434666726324292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,10240,0.021342222889264423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,768,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,10240,0.10297599765989517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,128,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,4096,0.3572026623619928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,3584,0.007105777661005656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,7168,0.08151555723614164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,32,0.00609422226746877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,8192,0.03447555502255758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,512,0.00509066672788726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,768,0.00610133343272739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,1024,0.004286222159862518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,1024,0.008124444219801161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,6144,0.011648000114493899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,6144,0.016522667474216886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,128,0.00609688874748018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,1024,0.012087999946541257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,6144,0.16189600361718073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,5120,0.005730666634109285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,12288,0.019869332512219746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,1536,0.006288888967699475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,128,0.006309333360857434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,2560,0.007350222104125553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,12288,0.06545333067576091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,5120,0.027500443988376196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,8192,0.021355556117163763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,128,0.011383111278216044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,3584,0.0165928883685006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,8192,0.014565333724021912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,5120,0.0243520008193122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,10240,0.012387555506494312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,256,0.006688888702127669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,2048,0.025411556164423626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,6144,0.03115644388728672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,256,0.014128888646761576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,7168,0.1409048901663886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,128,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,8192,0.011339555184046427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,256,0.0031191110610961914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,3072,0.016160888804329764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,16384,0.016132444143295288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,2048,0.02838133441077338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,2560,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,6144,0.009259555902745988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,512,0.010490666660997601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,4096,0.035777777433395386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,1024,0.0038613333470291565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,1024,0.006351111249791251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,1536,0.026927999324268762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,10240,0.04022044605678982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,65536,0.08374577760696411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,16384,0.016911110944218107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,4096,0.023248000277413264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,65536,0.04237066706021627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,16384,0.2764497862921821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,3584,0.01554844445652432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,256,0.01755822201569875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,3584,0.1990657779905531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,2560,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,4096,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,1536,0.006569777925809224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,64,0.0052062223354975385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,2560,0.007000000112586551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,6144,0.012896888785892062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,256,0.009304000271691216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,3584,0.01351288871632682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,65536,0.0806808869043986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,10240,0.016922665966881644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,768,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,128,0.016170667277442086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,2560,0.014153778553009033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,8192,0.015610666738616096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,3072,0.00629333323902554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,1536,0.010848888920413123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,2048,0.0063991112013657885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,5120,0.04914933443069458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,2048,0.015447111593352424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,1024,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,6144,0.1476915544933743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,65536,0.06225422355863783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,64,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,3072,0.007028444773621029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,6144,0.014926221635606555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,6144,0.17145866817898223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,1536,0.012039110892348819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,4096,0.10693333546320598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,2048,0.0069146665434042616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,3072,0.02283199959331089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,8192,0.07423910829755995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,1536,0.006213333457708359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,128,0.009206222163306342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,10240,0.03792266713248359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,256,0.003567999849716822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,1024,0.006137777947717243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,64,0.005554666535721884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,1536,0.009729777773221334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,5120,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,16384,0.05460266603363884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,12288,0.027958220905727808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,3072,0.007365333537260692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,16384,0.023157333334287006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,256,0.0074888889988263445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,1536,0.006664888726340399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,6144,0.06476622157626681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,256,0.0029102222373088202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,3584,0.02023555503951179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,256,0.008662222160233391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,2560,0.008355555435021719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,512,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,512,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,4096,0.01370133293999566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,2560,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,2048,0.0069715554515520734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,768,0.0042444442709287005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,768,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,2048,0.029101332028706867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,10240,2.96943367852105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,128,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,128,0.009719111025333405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,256,0.03169066707293192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,65536,0.09353955586751302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,512,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,2560,0.012061333490742577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,10240,0.015594666202863058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,1024,0.012097777591811286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,4096,0.014661333627170987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,2560,0.030992888742023047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,4096,0.01945244438118405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,128,0.009703111317422655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,10240,0.10043822394476996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,4096,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,2560,0.04461333486768934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,4096,0.01479822231663598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,3072,0.037447998921076454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,768,0.010072000324726105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,512,0.005657777604129579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,3584,0.10988355345196194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,768,0.015499555402331881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,1024,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,128,0.02646044393380483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,2560,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,6144,0.02420888841152191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,7168,0.009463110731707679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,512,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,7168,0.010671111444632212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,5120,0.01276177747382058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,8192,0.02628177735540602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,10240,0.03240799903869629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,10240,0.01351199961370892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,2048,0.01909244391653273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,5120,0.012416889270146688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,6144,0.06419822242524889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,512,0.005261333452330695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,12288,0.01460088955031501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,1536,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,1536,0.0069022224181228215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,8192,0.02294933299223582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,1024,0.016156444946924847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,64,0.00318755561278926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,1536,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,8192,0.011335999601417117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,32,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,512,0.03716977768474155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,10240,0.03445511062939962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,8192,0.010758221977286868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,512,0.01722311145729489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,3584,0.007229333122571309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,65536,0.6906515757242838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,8192,0.03381866547796462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,128,0.026478222674793665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,12288,0.02887466549873352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,1024,0.021614222062958613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,6144,0.009693333672152625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,5120,0.008887111312813228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,8192,0.013763555222087435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,65536,0.04674666788842943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,2560,0.016203555795881484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,4096,0.007788444558779399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,10240,0.016578666037983365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,3072,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,10240,0.03386577632692125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,6144,0.011086222198274402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,12288,0.013459555804729462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,8192,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,1536,0.005325333525737126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,3584,0.024696888195143804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,5120,0.046217779318491616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,5120,0.008328888979223039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,2048,0.019317333896954853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,3584,0.024678222007221643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,768,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,256,0.0069297779765393995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,1024,0.012727110750145383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,512,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,4096,0.013412444127930535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,8192,0.03648799988958571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,256,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,5120,0.009161777794361115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,6144,0.02926666537920634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,128,0.015508444772826301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,1024,0.016176889340082806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,65536,0.08537333541446263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,128,0.0034711110509104202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,3584,0.026039999392297532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,4096,0.08621244298087226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,2560,0.007311999797821045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,12288,0.006232888748248418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,2048,0.005992888990375731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,768,0.016170667277442086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,3584,0.0075866662793689305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,6144,0.025379555092917547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,512,0.005218666460778978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,5120,0.03778577844301859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,128,0.0052426668504873914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,2048,0.019135110908084445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,65536,0.06352088848749797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,8192,0.20415200127495658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,65536,0.06995733578999837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,768,0.011035555766688453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,512,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,256,0.004873777843183941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,4096,0.016346666547987197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,2048,0.020756444997257657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,512,0.011728888584507836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,2560,0.015840000576443143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,3584,0.008343111309740279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,16384,0.3798364533318414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,512,0.005509333478079901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,3072,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,16384,0.1629528866873847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,65536,0.13470310635036892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,6144,0.027054222093688116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,1024,0.004592000196377437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,2048,0.0536008874575297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,1024,0.0325617790222168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,65536,0.06318666537602742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,1536,0.007511110769377814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,3072,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,5120,0.01705777810679542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,768,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,4096,0.02056799994574653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,32,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,5120,0.008652444514963362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,6144,0.01218488895230823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,6144,0.01126488877667321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,2560,0.007280000381999546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,4096,0.01147377739350001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,2560,0.007355555891990662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,2560,0.015470221638679504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,1536,0.00996800015370051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,3584,0.016177778442700703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,5120,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,4096,0.012462221913867526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,2048,0.015461333923869662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,3072,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,32,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,1536,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,65536,0.3388604323069255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,3584,0.008299555215570662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,3584,0.0162124451663759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,2560,0.010782221953074137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,768,0.020297777321603563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,16384,0.023399111297395494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,1536,0.00795911086930169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,10240,0.012533333566453723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,1536,0.008976889153321585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,2048,0.019497777024904888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,8192,0.018400889303949144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,256,0.012448000411192576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,3072,0.08117510875066121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,1024,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,7168,0.010728889041476779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,2048,0.0166631109184689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,2560,0.007275555696752336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,8192,0.045463999112447105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,128,0.01483022173245748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,256,0.0032177778581778207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,65536,0.20177688863542345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,1536,0.007043555378913879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,512,0.005542222410440445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,1024,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,8192,0.019221333993805777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,16384,0.021280889709790547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,12288,0.06567022535536024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,65536,0.08904355764389038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,1536,0.09168089098400539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,256,0.005311111195219888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,128,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,6144,0.017608000172509086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,2048,0.015103111664454142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,7168,0.0768408907784356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,12288,0.03207644489076402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,1536,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,8192,0.0336151123046875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,4096,0.00977600034740236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,512,0.014499555031458536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,16384,0.023063111636373732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,8192,0.014091556270917257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,512,0.00630666646692488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,2560,0.01791733337773217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,768,0.00599911105301645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,3072,0.015495111544926962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,768,0.009932444327407414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,256,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,2048,0.014430221584108142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,5120,0.00795911086930169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,12288,0.013400889105266996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,12288,0.2868559890323215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,128,0.014111111561457316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,7168,0.04066666629579332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,4096,0.058635552724202476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,5120,0.018020444446139865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,4096,0.008033778104517195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,16384,0.06475733386145698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,3584,0.010392888552612728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,2048,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,3584,0.023623999622133043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,6144,0.009369778136412302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,3584,0.10404444403118557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,1536,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,3584,0.02680444386270311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,5120,0.046738667620552905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,7168,0.02032888929049174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,64,0.0029084444459941653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,8192,0.05460800064934624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,3072,0.019036443697081674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,2560,0.008654221892356873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,16384,0.03267200125588311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,512,0.006069333189063602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,65536,0.07305333349439833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,128,0.014422222971916199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,3072,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,65536,0.0826524429851108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,6144,0.020788444413079154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,1024,0.004237333519591225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,10240,0.02200444373819563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,12288,0.01276177747382058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,6144,0.017470222380426195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,32,0.0026071110947264563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,12288,0.09415377510918511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,768,0.022264000442292955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,1024,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,3072,0.017163554827372234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,768,0.01578044394652049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,3584,0.023955555425749883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,1536,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,65536,0.04705422123273214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,2560,0.044080889887279935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,2048,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,512,0.02130488885773553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,2048,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,512,0.006888000087605582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,512,0.003160888950030009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,16384,0.05807377894719442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,512,0.0033484444850020935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,16384,0.010919999745157031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,3072,0.06670488913853963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,6144,0.017239999439981248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,3072,0.016504888733228047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,768,0.044122666120529175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,512,0.003181333343187968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,12288,0.03228177626927694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,3072,0.01723022262255351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,3584,0.010794666906197866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,128,0.010022222167915767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,4096,0.028918223248587713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,10240,0.028879112667507593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,3584,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,3584,0.010096888575288985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,256,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,5120,0.009720000127951304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,5120,0.034725334909227155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,12288,0.11330310503641765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,2560,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,32,0.0029111111329661477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,512,0.014463999205165438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,3072,0.013459555804729462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,32,0.007313778003056844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,6144,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,3072,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,5120,0.011624000138706632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,32,0.003212444484233856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,5120,0.03614577651023865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,768,0.009992000129487779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,32,0.01276711126168569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,64,0.0028817777832349143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,2048,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,512,0.003858666867017746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,768,0.017528888252046373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,65536,0.05664177735646566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,5120,0.07159555620617337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,3072,0.025047999289300706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,7168,0.006633777585294511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,4096,0.012844444149070315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,16384,0.02093422247303857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,128,0.005208888815508948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,8192,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,1024,0.02628266645802392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,5120,0.040947556495666504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,3072,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,5120,0.021736888421906367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,10240,0.007931555310885111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,768,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,256,0.0074337778819931885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,1536,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,1024,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,12288,0.030682666434182063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,4096,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,10240,0.02366666661368476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,4096,0.023589332898457844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,768,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,10240,0.038358221451441445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,3072,0.009672889278994666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,2048,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,5120,0.008267555799749162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,32,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,512,0.01275466630856196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,1024,0.007667555577225155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,10240,0.013818666338920593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,3072,0.024107555548350017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,6144,0.012994666894276937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,5120,0.009682666924264696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,512,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,512,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,3584,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,2560,0.015831111205948722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,16384,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,16384,0.05734755595525106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,7168,0.014375110467274984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,128,0.005223999834722943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,1536,0.042697777350743614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,128,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,65536,0.05398044321272108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,32,0.0035217776894569397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,512,0.0038417776425679526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,5120,0.008627555436558193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,6144,0.011117333339320289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,3072,0.04409511221779717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,3584,0.007260444263617198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,7168,0.02548888822396596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,5120,0.012105777859687805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,12288,0.02050044470363193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,4096,0.107951111263699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,1536,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,1024,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,3584,0.01723022262255351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,6144,0.09281511439217462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,1536,0.006152888966931238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,5120,0.03197244471973843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,65536,0.0588631100124783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,768,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,256,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,2560,0.017847110827763874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,128,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,5120,0.010411555568377176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,768,0.02422133253680335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,1024,0.01772444446881612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,3072,0.042466666963365346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,6144,0.028852442900339764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,512,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,1536,0.01648622254530589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,16384,0.0457822216881646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,3584,0.0070426662762959795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,3584,0.020257777637905545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,7168,0.02223199937078688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,32,0.00630666646692488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,512,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,16384,0.02492800023820665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,128,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,1024,0.027368000812000696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,2048,0.008672888908121321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,12288,0.038145777251985334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,3584,0.017422222428851657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,7168,0.01891999940077464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,65536,3.2353591918945312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,128,0.02351644469632043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,5120,0.022355554832352534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,3584,0.017350221673647564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,65536,0.03246488836076524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,16384,0.023240889112154644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,256,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,2560,0.023403555154800415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,5120,0.5735271241929796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,3584,0.026206221845414903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,4096,0.007928888830873702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,7168,0.02104355560408698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,3584,0.048487110270394214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,2560,0.007294221884674496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,8192,0.024914667010307312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,256,0.005212444398138259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,6144,0.006500444478458828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,3584,0.018295110927687753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,256,0.003297777846455574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,128,0.0058240000572469496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,3584,0.007297777467303806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,5120,0.009763555394278632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,3072,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,32,0.002918222298224767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,4096,0.008775110873911116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,2560,0.007276444799370236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,3072,0.007818666597207388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,256,0.012876444392734103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,768,0.021827555365032617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,4096,0.04541511005825467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,4096,0.00927022182279163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,128,0.013972444666756524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,64,0.005992888990375731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,4096,0.00904444439543618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,12288,0.09747999906539917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,7168,0.016984888248973422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,5120,0.06859733661015828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,5120,0.015984889533784654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,64,0.004378666480382283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,256,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,6144,0.011031111081441244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,12288,0.02573066618707445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,256,0.015105777316623263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,768,0.01718133356836107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,256,0.008008889026112026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,5120,0.023631110787391663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,32,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,2560,0.04174222217665779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,1536,0.008658666577604082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,12288,0.013247110777431063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,1536,0.006593777901596493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,5120,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,65536,0.05656177467770047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,16384,0.015160888433456421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,12288,0.02792177928818597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,4096,0.034741332133611046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,256,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,8192,0.024680000212457445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,12288,0.04129955503675673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,4096,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,4096,0.019959999455346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,65536,0.0697502228948805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,1024,0.12252710925208198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,65536,0.06497510936525133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,1536,0.00665244460105896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,1024,0.032447109619776406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,10240,0.029883556895785864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,768,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,3072,0.01646222174167633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,12288,0.03877777854601542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,3072,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,1536,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,7168,0.023972445064120825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,256,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,7168,0.01722044414944119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,1536,0.004912888838185204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,65536,0.09201600154240926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,512,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,65536,0.06266666783226861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,5120,0.030807998445298936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,10240,0.01682400041156345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,128,0.005998222364319696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,5120,0.023481777972645227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,10240,0.0693217780854967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,7168,3.6817529466417103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,128,0.011692444483439127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,2048,0.017824000782436795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,5120,0.04325244492954678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,128,0.025032000409232244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,16384,0.020263110597928364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,65536,0.061943113803863525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,256,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,10240,0.012770666844315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,128,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,3584,0.016189333465364244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,128,0.01032977799574534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,3584,0.011355555719799466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,6144,0.022554667459593877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,6144,0.015628443823920358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,1024,0.018350222044520907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,512,0.014831999937693277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,2560,0.015498666299713982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,12288,0.13595733377668592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,16384,0.02083200050724877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,64,0.16694488790300158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,768,0.01647822227742937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,2048,0.005240889059172736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,3072,0.0074879998962084455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,256,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,6144,0.01242844429281023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,12288,0.0644764436615838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,3584,0.01036622209681405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,12288,0.05660533242755466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,512,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,12288,0.013064000341627332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,64,0.0028213332924577924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,3584,0.011180444724029966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,8192,0.0859066645304362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,12288,0.029151111841201782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,3072,0.030039999220106337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,1024,0.008295111358165741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,6144,0.00906133320596483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,2560,0.010371555884679159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,7168,0.019683554768562317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,256,0.006776000062624614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,128,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,5120,0.024060444699393377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,2048,0.02627022233274248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,2560,0.021382222572962444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,1536,0.014797333214018079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,2560,0.02342844506104787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,4096,0.014486221803559197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,16384,0.030926220946841772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,4096,0.024717332588301763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,65536,0.04351910948753357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,3072,0.2894551224178738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,1024,0.020231111182106864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,128,0.005292444593376583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,2048,0.006670222100284364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,2048,0.016152888536453247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,4096,0.06408711274464925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,4096,0.022322666313913133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,2560,0.008978666530715095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,10240,0.01625866691271464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,128,0.01462933255566491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,128,0.015802666544914246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,1024,0.07700266440709432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,128,0.011376000112957425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,6144,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,4096,0.018135999639829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,65536,0.07273511091868083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,256,0.005331555588377847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,12288,0.013571555415789286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,7168,0.10702755716111924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,2560,0.07802310917112562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,256,0.04213066564665901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,2048,0.006950221955776215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,7168,0.010006222460005019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,2048,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,3584,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,6144,0.03489244315359328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,65536,0.40916532940334743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,256,0.003569777641031477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,5120,0.009289777941173976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,8192,0.010374222364690568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,7168,0.014156444205178155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,64,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,65536,0.14486132727728948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,5120,0.02793955471780565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,4096,0.007912889122962952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,256,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,512,0.005561777700980504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,512,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,5120,0.011507555842399597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,512,0.019529778096410964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,8192,0.0306328899330563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,512,0.05121155579884847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,16384,0.03122666809293959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,3584,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,3072,0.08409333229064941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,768,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,2048,0.031088001198238794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,4096,0.008656889200210571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,3072,0.010058666268984476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,1024,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,1024,0.0207342223988639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,512,0.011040888726711273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,512,0.030356443590588037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,4096,0.015390222271283468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,8192,0.1904275549782647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,1024,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,2560,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,256,0.00517511119445165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,1536,0.00453244439429707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,1024,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,64,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,3072,0.007335111498832703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,1024,0.004559999952713649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,10240,0.014502222339312235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,1536,0.06877066691716512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,1536,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,128,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,1536,0.022470222579108343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,3584,0.012888000243239932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,1024,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,1536,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,768,0.0041804446114434134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,8192,0.015142222245534262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,4096,0.02299555473857456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,6144,0.1587511168585883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,2560,0.03657066822052002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,6144,0.0691137777434455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,2560,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,1024,0.006252444452709622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,12288,0.01311822235584259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,8192,0.018567111757066514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,32,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,32,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,512,0.014122666584120857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,1024,0.0041457778877682155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,768,0.013422222601042854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,512,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,8192,0.01461155547036065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,5120,0.01831822262869941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,10240,0.022639999787012737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,512,0.01309511148267322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,7168,0.04236977630191379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,512,0.017553778158293832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,2560,0.012134222520722283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,2048,0.01125511113140318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,8192,0.028580443726645574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,2560,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,3584,0.04443022277620104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,8192,0.024783111280865137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,65536,0.46189244588216144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,65536,0.07742844687567817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,8192,0.006575999988449945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,16384,0.03037422233157688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,32,0.006288888967699475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,256,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,32,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,7168,0.013269333375824822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,768,0.02919911013709174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,2048,0.032215111785464816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,3072,0.05268266797065735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,256,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,256,0.011142222417725457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,7168,0.016908443636364408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,7168,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,7168,0.009360888765917884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,3584,0.00924711094962226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,7168,0.07797955804400973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,256,0.05090399914317661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,6144,0.012078222301271228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,5120,0.011400000088744693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,7168,0.05835377507739597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,7168,0.05192800031767952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,256,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,2560,0.038745777474509344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,7168,0.024448888169394598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,8192,0.015601777368121676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,128,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,7168,0.00832177781396442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,16384,0.03021777669588725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,3072,0.006963555183675554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,12288,0.021021333005693223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,7168,0.026141333911154006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,12288,0.0730817781554328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,512,0.009843555589516958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,2560,0.039451556073294744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,16384,0.4015226629045274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,5120,0.02147644427087572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,8192,0.014161777165200977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,256,0.013099555340078143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,3584,0.01680799987581041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,512,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,16384,0.01313688854376475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,3584,0.010518222219414182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,8192,0.010699555277824402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,6144,0.0852853324678209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,8192,0.02161600026819441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,768,0.005940444353553984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,3584,0.007624000310897827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,32,0.0057591112951437635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,10240,0.02678933408525255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,12288,0.04712000158098009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,6144,0.00962488849957784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,8192,0.033568888902664185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,768,0.005883555445406172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,10240,0.022234666678640578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,2560,0.026172444224357605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,4096,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,4096,0.007975999679830339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,128,0.003485333174467087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,8192,0.024562666813532513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,256,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,128,0.021936888496081035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,4096,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,10240,0.02130399975511763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,8192,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,6144,0.024033778243594702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,16384,0.011083555718262991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,6144,1.2782026926676433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,2560,0.0069644442862934535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,12288,0.022672888305452134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,512,0.00526577772365676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,4096,0.6763511233859592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,32,0.004228444563017951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,128,0.02846133377816942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,32,0.004591111093759537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,6144,0.030805332793129817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,6144,0.056410663657718234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,4096,0.1299591064453125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,8192,0.01978755493958791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,1024,0.007820444802443186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,3072,0.012055111428101858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,6144,0.009364444348547194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,1536,0.029044446018007066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,3584,0.007941332956155142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,3584,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,12288,0.008278222547637092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,12288,0.037540445725123085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,12288,0.020462221569485135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,16384,0.02034311162100898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,2048,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,4096,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,7168,0.011034666664070554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,8192,0.022124444444974262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,4096,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,8192,0.07184355788760714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,8192,0.016861332787407767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,2560,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,5120,0.08394400278727214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,6144,0.013597333596812354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,3584,0.014768888552983602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,7168,0.012641777594884237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,3584,0.023174222972657945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,768,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,12288,0.030717333157857258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,32,0.009315555294354757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,768,0.004168888760937585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,10240,0.10766755872302586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,16384,0.07995644542906019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,6144,0.015136889285511441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,1536,0.008336889247099558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,768,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,512,0.012089778151777057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,8192,0.015606222881211175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,6144,0.02013688948419359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,7168,0.013767110804716745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,128,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,6144,0.009311111436949836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,7168,0.013415111435784234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,6144,0.02126755482620663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,1536,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,2560,0.015562666787041558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,7168,0.013906665974193148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,16384,0.10592622227138943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,1024,0.007111111448870764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,2048,0.008615111311276754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,4096,0.018271999226676095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,3072,0.009296889106432596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,65536,0.08651910887824164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,5120,0.008634666601816813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,768,0.008977777428097194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,16384,0.022016000416543748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,256,0.01514400045077006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,2048,0.027454222242037456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,4096,0.01351199961370892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,7168,0.014535110857751636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,8192,0.1255306667751736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,16384,0.0713662240240309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,8192,0.02149155570401086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,5120,0.011759999725553723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,1536,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,3072,0.08137866523530748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,10240,0.24958843655056426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,256,0.015173332558737861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,512,0.003236444460021125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,6144,0.01719377769364251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,5120,0.009017777939637503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,1024,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,128,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,10240,0.22156266371409097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,65536,0.05558222532272339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,8192,0.032680888970692955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,1024,0.016490666402710807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,16384,0.020230222079488967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,512,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,12288,0.032023111979166664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,5120,0.01405244403415256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,3584,0.010802666346232096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,4096,0.013929777675204806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,2560,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,6144,0.024680889315075342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,12288,0.02755911151568095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,1536,0.013790222505728403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,12288,0.03466399841838413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,7168,0.008601778083377415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,12288,2.3386141459147134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,1024,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,16384,0.036841776635911726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,65536,0.06222755379147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,1024,0.01420177850458357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,7168,0.0218240006102456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,3584,0.009872888525327047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,4096,0.009724444813198512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,12288,0.01961511042382982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,65536,0.12341511249542236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,65536,0.06060533391104805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,10240,0.04628088739183214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,1024,0.016861332787407767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,10240,0.02201777696609497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,512,0.0039048890272776284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,1024,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,7168,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,8192,0.022749332918061152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,1024,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,3072,0.009699555734793345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,1024,0.014160888062583076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,6144,0.02053244411945343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,3072,0.033815112378862165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,7168,0.02883466747072008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,5120,0.013078221844302284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,8192,0.26568177011277944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,4096,0.008303110798199972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,1024,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,6144,0.008676444490750631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,2048,0.016885333591037326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,64,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,65536,0.07150489091873169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,3584,0.008175111479229396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,3584,0.13314667012956408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,768,0.016162667009565566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,10240,0.45634667078653973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,1024,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,512,0.003823111040724648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,128,0.0031537777847713898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,12288,0.03710844450526767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,512,0.004587555511130227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,3584,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,32,0.0028728888266616394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,7168,0.04000622365209792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,512,0.014129777749379476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,2560,0.007644444704055786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,2048,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,7168,0.01740622189309862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,2560,0.013479111095269522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,1536,0.01662400033738878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,4096,0.007718222008811102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,4096,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,512,0.0083137775460879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,5120,0.016928889685206942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,6144,0.018552000323931377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,7168,0.02852177619934082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,4096,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,5120,0.01688799924320645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,3072,0.014130666851997375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,8192,0.014815110299322339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,32,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,768,0.015102222561836243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,4096,0.017125333348910015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,4096,0.01640888883007897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,256,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,8192,0.022387555903858606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,10240,0.27061155107286244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,10240,0.01205244412024816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,2560,0.025978666212823655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,1536,0.03520088725619846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,1536,0.01696266730626424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,7168,0.008171555896600088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,1536,0.01519377695189582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,2560,0.028049776951471966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,64,0.002773333340883255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,2560,0.4653609063890245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,1024,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,2560,0.007831999825106727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,3072,0.009330666727489894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,512,0.01274577776590983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,3072,0.020946666598320007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,4096,0.009719111025333405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,4096,0.011148444480366178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,512,0.0038293335172865125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,512,0.006753777878151999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,256,0.016531555189026725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,512,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,32,0.04188444548183017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,16384,0.009286222358544668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,10240,0.028451555305057104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,256,0.04894311229387919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,3072,0.022656889425383672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,768,0.007108444141017065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,12288,0.00904444439543618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,1024,0.06131111250983345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,512,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,1024,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,768,0.015081778168678284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,32,0.0028115556471877625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,7168,0.011210666762457954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,1536,0.004551110996140374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,12288,0.04203822215398153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,1536,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,3584,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,3072,0.024495111571417913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,5120,0.05857955747180515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,32,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,7168,0.02477866742346022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,128,0.011759111450778114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,64,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,12288,0.03043466806411743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,6144,0.009768000079525841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,4096,0.03908711009555393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,768,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,768,0.014144000079896716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,1536,0.006263999889294307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,768,0.01068800025516086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,12288,0.12842488288879395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,1024,0.0052355556852287715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,65536,0.049703111251195274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,1024,0.006484444356626934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,2560,0.07072799735599093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,128,0.005299555758635203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,65536,0.08307555649015638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,768,0.040235555834240384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,3584,0.009480000370078618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,6144,0.009266667068004608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,2048,0.028411553965674505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,10240,0.04814755585458544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,2560,0.011577777564525604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,4096,0.01053422192732493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,64,0.007799111306667328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,6144,0.06629688872231378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,256,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,5120,0.02056799994574653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,3584,0.015271999769740634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,768,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,4096,0.01757600075668759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,256,0.005516444643338521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,3072,0.007713778151406183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,6144,0.02919644448492262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,1024,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,8192,0.016356445021099515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,7168,0.015458666616015963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,1024,0.010080888867378235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,2560,0.006461333483457565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,16384,0.07739022043016222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,128,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,512,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,65536,0.02436622149414486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,65536,0.09079821904500325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,7168,0.01385600037044949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,1536,0.01790844400723775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,1024,0.4576711124844021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,12288,0.02955199943648444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,3072,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,6144,0.012952889005343119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,1536,0.016167999969588388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,5120,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,65536,0.14324622684054905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,2560,0.012033777932325998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,3584,0.01498933302031623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,1536,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,12288,0.044101334280437894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,4096,0.07455999983681573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,6144,0.017875555488798354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,5120,0.035102221700880266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,64,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,65536,0.07124444511201647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,6144,0.03180355495876736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,7168,0.009687110781669617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,1024,0.008669333325492011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,768,0.003863999827040566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,7168,0.010024000373151567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,7168,0.017845332622528076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,2560,0.016840888394249808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,2048,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,256,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,10240,0.01644355555375417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,256,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,3072,0.016161777906947665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,6144,0.009372444616423713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,12288,0.0358168880144755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,5120,0.057560000154707164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,64,0.0939191116227044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,6144,0.05386044581731161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,2560,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,16384,2.062865787082248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,1024,0.012686221963829465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,65536,0.08156266477372912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,12288,0.009666666388511658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,4096,0.01829777823554145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,16384,0.023045332895384893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,8192,0.025996444953812495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,1536,0.004579555657174853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,6144,0.008928889201747047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,64,0.010688888529936472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,65536,0.23473776711357963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,256,0.016203555795881484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,5120,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,3072,0.01722755531469981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,10240,0.0273644444015291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,512,0.003549333247873518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,12288,0.08357155323028564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,768,0.008918222453859117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,3584,0.04686311218473646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,16384,0.03204177816708883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,512,0.007698666718271043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,32,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,10240,0.01613333324591319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,256,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,128,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,32,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,12288,0.029432889488008287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,8192,0.023927110764715407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,2560,0.014846222268210517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,1024,0.00424888895617591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,10240,0.05609689156214396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,8192,0.047819554805755615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,3072,0.014887111054526435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,8192,0.044998221927218966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,1024,0.023194667365815904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,3584,0.01108088923825158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,5120,0.015656888484954834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,8192,0.0378115541405148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,3584,0.03460888730155097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,8192,0.011011555790901184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,512,0.003499555504984326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,10240,0.018367111682891846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,65536,0.04646044307284885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,256,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,3584,0.034148444732030235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,1024,0.012614222036467658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,3072,0.009014222357008193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,64,0.00582311095462905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,1024,0.01517955627706316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,512,0.016826666063732572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,768,0.003567999849716822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,65536,0.06544088655047946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,10240,0.016063110695944894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,8192,0.044813334941864014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,1536,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,64,0.01550044450494978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,65536,0.018585777944988675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,8192,0.01350311107105679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,256,0.005186666631036335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,5120,0.03140888942612542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,1536,0.006786666810512543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,12288,0.03665333324008518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,12288,0.013163555827405719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,64,0.0028799999919202593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,3072,0.03414933217896356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,65536,0.12056977219051784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,5120,0.017861333158281114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,3584,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,3072,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,65536,0.047945777575174965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,256,0.01757244434621599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,768,0.005690666536490123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,128,0.009053332938088311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,512,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,3072,0.00794400026400884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,6144,0.011207111179828644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,4096,0.009777777724795872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,6144,0.013166222307417126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,3584,0.02786133355564541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,1536,0.010661332971519895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,512,0.006217777729034424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,7168,0.06623911195331149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,1536,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,256,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,2048,0.006414222220579783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,4096,0.015072888798183866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,1024,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,4096,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,32,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,6144,0.017264889346228707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,3072,0.027136888768937852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,1024,0.02494844463136461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,8192,0.18254310554928252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,5120,0.011545778148704104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,65536,0.49435822168986004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,16384,0.031522668070263334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,12288,0.03091644578509861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,6144,0.027330666780471802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,256,0.014893333117167154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,64,0.005889777921968036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,65536,0.015558222929636637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,5120,0.018626666731304593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,4096,0.00796622203456031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,8192,0.0239982224173016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,7168,0.009727111293209923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,5120,0.009007111191749573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,32,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,32,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,1024,0.06356444623735216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,512,0.013063111239009433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,5120,0.04761688907941183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,32,0.002765333279967308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,3584,0.017522666189405654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,2048,0.008468444148699442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,2560,0.00962488849957784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,256,0.004212444441186057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,768,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,3584,0.0414613319767846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,8192,0.03246577911906772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,10240,0.01883555617597368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,768,0.013827555709415011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,2560,0.006899555524190267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,6144,0.011471110913488599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,5120,0.024321777953041926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,8192,0.010386666489972008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,2560,0.01477777792347802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,256,0.03456977672047085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,8192,0.031203554736243352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,4096,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,6144,0.009357333183288574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,8192,0.01956622302532196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,12288,0.09840889109505548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,6144,0.03788888785574172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,3072,0.00869422240389718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,16384,0.023952000670962866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,12288,0.029556443293889362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,6144,0.30180443657769096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,4096,0.027038221557935078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,128,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,7168,0.20083022117614746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,1024,0.06181510951783922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,16384,0.028192000256644353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,4096,0.02557777696185642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,7168,0.006054222169849608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,8192,0.01991911066903008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,2048,0.005370666583379109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,16384,0.044919110006756253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,4096,0.017166222135225933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,3584,0.006085333310895496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,768,0.003636444608370463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,10240,0.014139556222491793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,1536,0.008648888932334052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,10240,0.023285332653257582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,2048,0.018258665998776753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,4096,0.021963554951879714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,1024,0.008120888637171851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,1536,0.008063111040327284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,5120,0.9921226501464844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,4096,0.09457333220375909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,7168,0.02408711115519206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,7168,0.05781955851448906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,3072,0.007330666813585494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,1024,0.008363555702898238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,512,0.014473777678277759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,7168,0.021780444516075984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,3584,0.27990934583875865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,1536,0.02443466583887736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,10240,0.014187556174066333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,1536,0.05019288924005297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,2048,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,2560,0.007307555940416124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,32,0.0028853333658642242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,6144,0.018976888722843595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,3072,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,16384,0.03175910976197984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,2560,0.010219555762079027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,3584,0.008344888687133789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,768,0.016534222496880423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,2560,0.00868088917599784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,3072,0.025036444266637165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,7168,0.02715377840730879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,128,0.0031759999692440033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,2560,0.010296000374688042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,8192,0.011427555647161273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,7168,0.02926666537920634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,5120,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,768,0.017159110969967313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,2560,0.013753777576817406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,10240,0.020271999968422782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,3584,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,64,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,10240,0.020280000236299302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,1536,0.006611555400821898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,4096,0.01993688941001892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,1536,0.02218222286966112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,4096,0.019873778025309246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,2560,0.07891199986139934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,10240,0.027114666170544092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,256,0.013813333378897773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,8192,0.033414221472210355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,12288,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,3072,0.06477599673800997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,1536,0.009679999616410997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,1536,0.007670222057236566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,128,0.01553600033124288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,1536,0.011776000261306763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,3072,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,12288,0.014120888378885059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,5120,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,6144,0.01276177747382058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,64,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,3072,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,8192,0.012721777790122561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,12288,0.00870577742656072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,1536,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,32,0.006821333534187741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,2560,0.00852177788813909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,16384,0.040051556295818753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,1024,0.013438222308953604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,1024,0.004222222086456087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,3072,0.006992888947327931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,256,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,5120,0.00865955568022198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,7168,0.43691645728217227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,4096,0.355043543709649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,128,0.01887022289964888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,6144,0.021337777376174927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,512,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,256,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,1024,0.021703999903466966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,8192,0.012810666528013019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,7168,0.02176355487770504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,64,0.0031537777847713898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,512,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,3584,0.022647110952271357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,65536,0.10232533348931207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,1536,0.01827822294500139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,10240,0.018929777873886954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,128,0.01549600064754486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,10240,0.011688888900809817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,6144,0.059663110309176974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,1536,0.01573866605758667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,7168,0.010677333507272931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,2560,0.013691555294725629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,4096,0.01737511157989502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,8192,0.03622311022546556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,64,0.002971555623743269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,16384,0.149235553211636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,4096,0.007460444337791867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,256,0.006753777878151999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,128,0.0030986666679382324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,1024,0.007694222033023834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,256,0.01493866741657257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,5120,0.008978666530715095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,64,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,768,0.017283555534150865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,8192,0.01460088955031501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,1536,0.004612444589535396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,16384,0.07427199681599934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,2560,0.028346667687098186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,32,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,4096,0.011695110963450538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,128,0.012413333687517377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,7168,0.010296888649463654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,5120,0.02480711042881012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,65536,0.16611466142866346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,1024,0.01460533340771993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,8192,0.010671111444632212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,65536,0.13557510905795628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,12288,0.09432978100246853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,3584,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,512,0.01311466677321328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,512,0.01517511076397366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,10240,5.349868350558811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,16384,0.6974044375949435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,10240,0.05520711011356778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,7168,0.027730666928821143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,128,0.020453333854675293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,12288,0.7294053501553006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,128,0.02159999973244137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,768,0.014491554763582019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,4096,0.01514933341079288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,768,0.015506666567590503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,7168,0.016522667474216886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,32,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,128,0.0035217776894569397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,128,0.009768889182143742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,768,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,1024,0.007676444119877285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,16384,0.028392887777752344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,1024,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,10240,0.028805332051383123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,8192,0.14670222335391575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,3584,0.03902488946914673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,12288,0.02507466740078396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,65536,0.39773066838582355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,65536,0.15213599469926622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,256,0.005296000176005893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,5120,0.017255110873116385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,768,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,3584,0.009345778160625035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,7168,0.02873777680926853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,2048,0.015141333142916361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,3072,0.00847822262181176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,8192,0.03736799955368042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,1024,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,5120,0.00833511104186376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,2048,0.009696000152164036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,2048,0.016882666283183627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,16384,0.016823111308945548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,12288,0.083569778336419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,65536,0.06702578067779541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,2560,0.011776888536082374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,16384,0.37156444125705296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,2048,0.007941332956155142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,256,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,5120,0.011101333631409539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,3584,0.020928889513015747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,3072,0.013789333403110504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,4096,0.017881777551439073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,768,0.008348444269763099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,2560,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,3072,0.08127644326951769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,768,0.0038817777401871155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,2048,0.0351839992735121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,6144,0.2455831103854709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,10240,0.023065777288542852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,16384,2.0095272064208984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,16384,0.3478204409281413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,512,0.005231110999981563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,6144,0.04116888840993246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,3584,0.01705333259370592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,7168,0.011738667057620155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,1024,0.02786933382352193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,768,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,4096,0.00870311094654931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,64,0.006175111151403851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,768,0.0058319999112023254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,65536,0.11825422445933025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,12288,0.02075466679202186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,5120,0.008839999636014303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,4096,0.020032000210550096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,3072,0.011310222248236338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,65536,0.3004382186465793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,7168,0.012952889005343119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,768,0.01108888867828581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,16384,0.03292977809906006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,6144,0.007642666498819987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,2560,0.08484888739056057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,4096,0.006583999842405319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,12288,0.013432888521088494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,3584,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,16384,0.043531556924184166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,8192,0.010810666614108615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,3072,0.01000533335738712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,2048,0.005212444398138259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,3584,0.01791911158296797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,128,0.003148444410827425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,16384,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,256,0.006753777878151999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,2048,0.010011555420027839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,6144,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,10240,0.029501332177056208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,2560,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,768,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,12288,0.0610497792561849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,128,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,7168,0.019349333312776353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,8192,0.010808000134097205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,10240,0.01204533295498954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,16384,0.019692444139056735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,128,0.006232888748248418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,768,0.014893333117167154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,32,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,16384,0.2283093399471707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,4096,0.05489244394832187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,65536,0.0786551104651557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,256,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,65536,0.07486755318111844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,2560,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,3584,0.024821332759327356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,65536,0.0754257771703932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,32,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,10240,0.015666666958067153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,8192,0.09326755338244969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,2048,0.006555555595291986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,256,0.006624889042642381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,6144,0.4267920123206244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,2560,0.013767999907334646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,1024,0.007710222568776872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,65536,1.059575080871582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,6144,0.009316444396972656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,2048,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,2048,0.010508444574144153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,256,0.009738666315873465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,7168,0.04035822219318814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,10240,0.04581155710750156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,1024,0.00612088872326745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,1536,0.006231110956933763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,4096,0.008316444853941599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,768,0.028208888239330713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,10240,0.01789155602455139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,256,0.026683555708991155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,10240,0.14356799920399985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,8192,0.026183999247021143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,7168,0.018533334136009216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,4096,0.02683911058637831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,128,0.003087999920050303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,1536,0.023212444451120164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,10240,0.02050488856103685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,1024,0.020288000504175823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,6144,0.06881244315041436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,8192,0.010452444354693094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,7168,0.028195556667115953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,12288,0.04366311099794176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,3072,0.008008889026112026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,10240,0.035402665535608925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,6144,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,7168,0.032307555278142296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,10240,0.011443555355072021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,512,0.007165333463086023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,8192,0.11763111750284831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,4096,0.0707555545700921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,16384,0.03223199976815118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,1024,0.004528000122971005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,5120,0.01202044470442666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,2048,0.01130222198035982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,7168,0.014153778553009033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,128,0.0038168889780839286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,2048,0.01700355609258016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,3584,0.022135999467637803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,768,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,3072,0.00868000007337994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,65536,0.06130044327841865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,12288,0.01906044450071123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,3584,0.013270222478442721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,2048,0.015266666809717814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,256,0.01533066729704539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,32,0.0034044444974925783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,1536,0.008126222425036961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,8192,0.11356533898247613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,12288,0.018067555295096505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,7168,0.010504888991514841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,256,0.09791466924879287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,2048,0.009159999589125315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,1024,0.006360000206364526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,7168,0.01495822270711263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,256,0.0031946665710873077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,6144,0.03915822174814012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,7168,0.037384000089433454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,1024,0.014194667339324951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,2048,0.007245333658324347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,128,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,8192,0.009401777552233802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,3584,0.007270221908887227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,32,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,8192,0.050507555405298867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,32,0.0031715554909573663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,7168,0.018592000007629395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,6144,0.20462932851579455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,16384,0.34419645203484434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,12288,0.4395662148793538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,6144,0.09474755658043756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,64,0.006238222122192383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,2560,0.012768000364303589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,10240,0.015928000211715698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,768,0.010716444088353051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,10240,0.10746133327484131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,512,0.003559111307064692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,65536,0.07164000140296088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,2048,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,8192,0.030466665824254353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,2048,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,2048,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,10240,0.010823999842007955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,10240,0.018420444594489205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,1024,0.006024888820118374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,65536,0.06776089138454862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,5120,0.02071911096572876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,8192,0.008328888979223039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,6144,0.02970222300953335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,3584,0.01148888882663515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,65536,0.13114666938781738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,64,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,7168,0.23420977592468262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,512,0.01331555595000585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,2048,0.10327288839552139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,65536,0.0640595555305481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,5120,0.02315644423166911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,1024,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,768,0.015517334143320719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,2560,0.007356444166766272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,65536,0.6583937538994683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,6144,0.026707554856936138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,32,0.0032426667296224167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,16384,0.08025511105855306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,128,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,10240,0.025311999850802954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,6144,0.06342222293217976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,128,0.005868444426192178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,5120,0.012893333203262754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,512,0.01312622262371911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,3584,0.00794400026400884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,8192,1.9908667670355902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,65536,0.869659529791938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,2048,0.011181332998805575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,1024,0.013225778109497495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,1024,0.006021333237489064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,5120,0.009214222431182861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,1536,0.037949333588282265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,12288,0.018255111243989732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,6144,0.009082666701740688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,32,0.004957333207130432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,7168,0.010000000397364298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,2560,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,10240,0.07626755370034112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,10240,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,7168,0.03795555565092299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,1536,0.007667555577225155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,128,0.012084444363911947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,4096,0.03160888950030009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,4096,0.018809777167108323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,1024,0.01537333263291253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,64,0.003142222141226133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,3072,0.006741333338949416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,128,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,128,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,4096,0.15041155285305446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,512,0.008270222279760573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,6144,0.01831466621822781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,8192,0.022678222921159532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,256,0.003216888962520493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,5120,0.01444355481200748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,6144,0.018043556147151522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,2048,0.04546133346027798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,512,0.005526222288608551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,4096,0.007975999679830339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,256,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,5120,0.00904888908068339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,2560,0.010326222413116032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,2048,0.027056889401541814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,256,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,4096,0.011717333561844297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,128,0.011403555671374003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,65536,0.06614222129185994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,8192,0.019595555133289762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,4096,0.09687911139594184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,256,0.0032133333798911837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,16384,0.015860444969601102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,12288,0.05409333440992567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,4096,0.02327288852797614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,7168,0.04916800061861674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,6144,0.02366399930583106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,4096,0.021336000826623704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,8192,0.025462221768167283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,768,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,16384,0.014929778046078153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,2560,0.013725333743625216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,6144,0.01587022178702884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,12288,0.01314222233162986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,4096,0.022061333060264587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,5120,0.009043555292818282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,2560,0.011204444699817233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,128,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,1536,0.005494222044944763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,6144,0.009000889129108852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,3072,0.03614666726854112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,2048,0.011733333269755045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,256,0.00815644446346495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,10240,0.050091554721196495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,2560,0.043818665875328906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,32,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,10240,0.043306665288077466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,65536,0.11610844400193955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,64,0.009492444495360056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,5120,0.012201777762836881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,512,0.014104000396198697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,1024,0.004877333425813251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,128,0.005296000176005893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,512,0.004563555535342959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,1536,0.020136000381575692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,2560,0.00656444455186526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,6144,0.013027555412716336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,128,0.01720533271630605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,16384,0.030566222137875024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,3584,0.026271111435360376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,512,0.013423110875818463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,6144,0.01094044413831499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,65536,0.11138310697343613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,5120,0.4648310873243544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,3072,0.007330666813585494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,256,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,5120,0.028104000621371802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,768,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,2048,0.005323555734422471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,2048,0.012819555898507437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,1536,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,1536,0.007679999702506595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,3072,0.0144533332851198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,65536,0.12230043941073948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,2048,0.015185778339703878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,16384,0.00795022232664956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,10240,0.051926222112443715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,8192,0.017130666308932833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,65536,0.19773156113094756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,12288,0.027053332991070215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,2560,0.014110222458839417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,5120,0.026940445105234783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,2048,0.026546667019526165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,65536,0.12796000639597574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,3584,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,2560,0.005730666634109285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,10240,0.03012622065014309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,2560,0.014495111174053617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,1536,0.010761777559916178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,3584,0.017737777696715463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,5120,0.046077334218555026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,2048,0.013769778112570444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,2560,0.006676444576846228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,512,0.005603555589914322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,12288,0.04012177719010247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,65536,0.04518933428658379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,768,0.030780444542566936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,128,0.005640000104904175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,4096,0.008016889293988546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,12288,0.03113599949412876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,8192,0.03923644291030036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,128,0.006575111299753189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,1536,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,3584,0.08791555298699273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,16384,0.03551466597451104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,2048,0.017705778280893963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,4096,0.01686044368478987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,32,0.0028364445186323593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,12288,0.025991999440722998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,16384,0.05334666702482435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,1024,0.0061048890153567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,10240,0.07661955886416964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,256,0.016561778055297006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,128,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,1536,0.004545777622196409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,4096,0.01686044368478987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,6144,0.01205244412024816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,1024,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,3584,0.007610666255156199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,3584,0.012063999970753988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,4096,0.009673777553770278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,5120,0.014150222142537435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,2560,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,768,0.08902400069766575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,7168,0.013694222602579327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,3584,0.007662222617202335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,10240,0.05875377522574531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,4096,0.01754755609565311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,5120,0.009983110758993361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,8192,0.04713066750102573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,4096,0.010416888528399997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,8192,0.022287110487620037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,768,0.010552888943089379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,2560,0.012291555603345236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,512,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,10240,0.022808889547983806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,10240,0.009762666291660732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,128,0.006330666856633292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,12288,0.03713066710366143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,512,0.005986666513813867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,4096,0.03219733304447598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,10240,0.042745778958002724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,2048,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,768,0.0037439999481042228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,1536,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,64,0.0029173334025674393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,2048,0.0329146683216095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,1024,0.012392889294359418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,256,0.031390223238203264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,3584,0.007731555236710443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,512,0.013782222237851886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,3072,0.017250667015711468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,65536,0.05350844396485222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,1536,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,5120,0.008967999782827165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,6144,0.019703111714786954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,4096,0.009336888790130615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,768,0.07331822315851848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,12288,0.03018666638268365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,16384,0.044531553983688354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,4096,0.012587555580668978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,7168,0.014380445082982382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,256,0.00574933323595259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,512,0.003866666720973121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,8192,0.016399110356966656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,10240,0.010695111420419482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,1536,0.01793511046303643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,2048,0.009480888644854227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,3584,0.010739555789364709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,5120,0.008625778059164682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,65536,0.08352355824576484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,10240,0.023909333679411147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,4096,0.022798221972253587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,1024,0.014203555054134794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,1024,0.015438222222858004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,4096,0.01496444476975335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,1536,0.013769778112570444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,1536,0.026957333087921143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,256,0.0035137778355015647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,2048,0.023073777556419373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,2560,0.01722755531469981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,65536,0.06461777951982287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,4096,0.06213511361016167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,5120,0.029527111185921565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,1536,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,3584,0.008969777988062965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,8192,0.2698631021711561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,5120,0.017983999517228868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,65536,0.08044177956051297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,12288,0.052001777622434825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,7168,0.02437955637772878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,3072,0.011874666644467248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,3584,0.018238221605618794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,6144,0.015451555450757345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,4096,0.011133333047231039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,1024,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,512,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,7168,0.01759022143152025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,128,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,512,0.02198133369286855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,12288,0.033589333295822144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,128,0.00602311102880372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,10240,0.022389334109094407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,65536,0.1638008885913425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,7168,0.05616177452935112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,512,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,65536,0.06437244680192736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,7168,0.013404444687896304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,10240,0.018008889423476327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,7168,0.010007999837398529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,8192,0.007408888803588019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,64,0.0075911109646161394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,5120,0.017875555488798354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,3072,0.01514844430817498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,768,0.01517688896920946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,2048,0.006955555743641323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,512,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,12288,0.0200311111079322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,3072,0.01616533266173469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,8192,0.030142221185896132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,2560,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,65536,0.04999644557634989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,7168,0.01035733355416192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,5120,0.03658755620320638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,256,0.005990222096443176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,3072,0.010411555568377176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,10240,0.011755555868148804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,8192,0.031167109807332356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,3584,0.01762488815519545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,512,0.059152000480228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,128,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,3072,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,5120,0.016543999314308167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,16384,0.015512888630231222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,5120,0.053301334381103516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,512,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,256,0.002983999955985281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,1024,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,64,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,2048,0.012694222231705984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,65536,0.10957777500152588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,5120,0.12922844621870253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,256,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,5120,0.03515911102294922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,4096,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,7168,0.012379555238617791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,256,0.014146667387750415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,128,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,5120,0.016893333858913846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,4096,0.009078222016493479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,5120,0.07977688974804349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,10240,0.03331644336382548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,16384,0.01981599960062239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,1536,0.01722755531469981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,256,0.003240888938307762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,64,0.0035075553589397003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,6144,0.024358222881952923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,768,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,3584,0.011054221954610614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,256,0.03052177694108751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,65536,0.13789689540863037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,2048,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,6144,0.027293332748942908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,768,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,128,0.005583111196756363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,64,0.002520888836847411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,4096,0.022060443957646687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,512,0.0035573335157500375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,3584,0.01923377811908722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,12288,0.17966932720608184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,6144,0.013172444370057849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,10240,0.025569778349664476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,768,0.018194667167133756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,2560,0.029768887493345473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,2560,0.016173332929611206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,8192,0.008295999632941352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,6144,0.010336000058386061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,2560,0.008970666262838576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,6144,0.011027555498811932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,2048,0.021928000781271193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,7168,0.015840889679061044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,6144,0.009276444713274637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,1024,0.016723554995324876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,64,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,512,0.01722666621208191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,128,0.003167111012670729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,8192,0.010737777584128909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,8192,0.010489777558379702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,16384,0.20491200023227266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,8192,0.02599999970859951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,256,0.009689778089523315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,7168,0.02941688895225525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,7168,0.014276444911956787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,1536,0.013043555948469373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,1536,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,16384,0.050572444995244346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,16384,0.07162933217154609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,5120,0.013111111190583972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,6144,0.01998044384850396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,128,0.01513155632548862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,4096,0.007963555554548899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,128,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,3584,0.009156444834338294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,768,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,16384,0.023818666736284893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,5120,0.02588711182276408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,256,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,3584,0.008030222521887885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,8192,0.017624000708262127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,128,0.0053173332578606075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,1024,0.004490666505363253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,8192,0.027248000105222065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,4096,0.01274222218328052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,2048,0.0232933329211341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,16384,0.0674640006489224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,3072,0.09727733002768622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,3584,0.008672888908121321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,2048,0.020580444071027968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,12288,0.0127368892232577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,768,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,6144,0.019741333193249173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,8192,0.02676800058947669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,256,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,256,0.016184889607959323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,768,0.01686755485004849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,128,0.014785778191354541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,16384,0.017610667480362784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,10240,0.02403555479314592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,2048,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,10240,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,768,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,1536,0.08185244268841214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,3584,0.014484445254007975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,512,0.003495111233658261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,7168,0.011052444577217102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,10240,0.025172443853484258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,6144,0.025584000680181716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,8192,0.066557334529029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,65536,0.10748089022106594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,4096,0.11615467071533203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,2048,0.017157332764731515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,768,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,128,0.0029111111329661477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,3584,0.009684444301658207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,5120,0.013616888887352414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,128,0.013776889277829064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,2560,0.016187555260128446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,65536,0.0847475528717041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,8192,0.00831200016869439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,32,0.0028071111689011254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,8192,0.013175110850069257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,512,0.0052737775776121355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,512,0.01346933344999949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,1024,0.009898666706350114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,2560,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,65536,0.048862222168180675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,7168,0.13805866241455078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,65536,0.1410177813635932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,12288,0.013941332697868347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,7168,0.018211555149820115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,2048,0.02407466702991062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,16384,0.0670071111785041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,16384,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,7168,0.02895911203490363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,65536,0.12046578195359971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,8192,0.010737777584128909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,1024,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,3584,0.021877333521842957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,12288,0.03501244386037191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,256,0.0053173332578606075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,16384,0.007803555164072249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,2048,0.006362666686375936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,32,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,512,0.0032133333798911837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,7168,0.020382222202089097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,16384,0.016373333003785875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,3072,0.010413332945770687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,12288,0.5860631201002333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,64,0.018586667047606576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,65536,0.09973422023985122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,10240,0.02917155623435974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,3584,0.008248888783984715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,6144,0.011889778077602386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,5120,0.010999110837777456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,256,0.006881777611043718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,1536,0.012725333372751871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,10240,0.0339662233988444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,8192,0.08148355616463555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,5120,0.019337778290112812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,4096,0.011138666835096149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,3584,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,4096,0.010079111489984725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,65536,0.050071110328038536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,7168,0.41049332088894314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,16384,0.047839999198913574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,2048,0.009054222040706212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,3584,0.012098666694429187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,8192,0.018051554759343464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,2048,0.02606755495071411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,2560,0.02202400068442027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,3072,0.011020444333553314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,6144,0.010052444206343757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,2560,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,10240,0.012058667010731168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,768,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,1536,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,1024,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,8192,0.02492888934082455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,4096,0.07768799861272176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,10240,0.012395555774370829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,768,0.01907199952337477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,8192,0.012913777596420713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,5120,0.007334222396214803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,2048,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,1024,0.008124444219801161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,2560,0.014614222778214348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,768,0.0038657776183552216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,32,0.0028328889360030494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,1536,0.03083911206987169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,256,0.011559111376603445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,2560,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,65536,0.1553528838687473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,10240,0.01716444392999013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,2048,0.005333333379692501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,3584,0.18376178211636016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,4096,0.007807110746701558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,6144,0.05004533463054233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,32,0.0038559999730851916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,8192,0.06323377953635322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,128,0.004561777744028303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,2560,0.009127111070685917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,768,0.015122666954994202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,768,0.04295555419392056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,3584,0.007660444411966536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,6144,0.012763555679056378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,16384,0.027672888504134283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,3072,0.09720000293519761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,128,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,6144,0.019916445016860962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,1536,0.007256000406212277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,3072,0.007670222057236566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,768,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,16384,0.04650133185916477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,1024,0.003847111016511917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,768,0.03458222084575229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,12288,0.030125333203209773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,256,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,5120,0.011058666639857821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,6144,0.0123857781291008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,8192,0.014656888114081489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,10240,0.02545333405335744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,2560,0.007631999750932057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,4096,0.012073777616024017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,16384,0.019306666321224637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,3072,0.007716444631417592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,64,0.004569777597983678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,3584,0.012140444583363004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,4096,0.008036444584528605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,256,0.0031742221779293488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,5120,0.008653333617581261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,10240,0.09572622511121963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,3072,0.03269066744380527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,1536,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,3584,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,1536,0.02956177790959676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,10240,0.5780622164408366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,1024,0.004590222405062782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,7168,0.0159244438012441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,10240,0.03033866816096836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,65536,0.3393093215094672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,2048,0.009167110754383935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,512,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,1536,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,512,0.0031555555760860443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,128,0.015466666883892484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,6144,0.011439999772442712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,2560,0.0069057775868309875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,1024,0.010230222509966956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,7168,0.020240000552601285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,7168,0.031325333648257785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,10240,0.028392000330819026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,5120,0.04105777872933282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,12288,0.03529599971241421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,3072,0.021588444709777832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,2048,0.006996444529957241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,2048,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,10240,0.06088799900478787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,12288,0.015471110741297403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,3072,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,6144,0.017650667164060805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,2560,0.009353777600659264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,7168,0.04296711087226868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,6144,0.008992888861232335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,1024,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,64,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,16384,0.2085404396057129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,65536,0.08421599864959717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,6144,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,64,0.005190222213665645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,3584,0.016392888294325936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,4096,0.0463066664006975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,10240,0.026102221674389307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,4096,0.03485777642991807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,10240,0.4155697822570801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,128,0.015474667151769003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,10240,0.026865777042176988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,2048,0.010399111443095738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,16384,0.02000711030430264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,10240,1.2747679816351998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,1024,0.017875555488798354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,1536,0.011014222270912595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,16384,0.02405244443151686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,768,0.006659555352396435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,768,0.04307111104329427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,256,0.011791999969217511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,3072,0.05283288823233711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,1024,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,1536,0.005803555664088991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,5120,0.04415466719203525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,256,0.002983111060327954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,12288,0.020875554945733812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,5120,0.03524355424775018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,12288,0.020423110988405015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,256,0.005881777654091517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,16384,0.013028444515334235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,256,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,2560,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,2048,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,12288,0.01308444473478529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,512,0.013463111387358772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,128,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,3584,0.14450844128926596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,4096,0.008373333348168267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,128,0.006508444332414203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,2560,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,12288,0.029712888929578993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,1024,0.06433955828348796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,512,0.031094223260879517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,256,0.006153777655627992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,128,0.014800889624489678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,1536,0.004911111046870549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,12288,0.008295111358165741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,3072,0.01555288831392924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,2560,0.02812533246146308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,512,0.0042035554846127825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,3584,0.02693866689999898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,2048,0.028896888097127277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,1536,0.04794755578041077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,12288,0.08493955267800225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,768,0.013553777502642738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,16384,0.06769777668846978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,10240,0.023502222365803186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,128,0.003590222034189436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,256,0.01069155583779017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,128,0.014839111102951897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,65536,1.021791140238444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,1024,0.04412710997793409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,256,0.023359111613697473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,1536,0.019991111424234178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,10240,0.02678488857216305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,6144,0.03014844324853685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,256,0.015824000040690105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,2560,0.06480355395211114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,16384,0.17797422409057617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,12288,0.02502222193611993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,8192,0.04160533348719279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,10240,0.021945777866575453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,5120,0.1321457756890191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,16384,0.030053334103690252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,10240,0.07224355803595649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,2560,0.017588444881969027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,1536,0.014488889111412896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,3072,0.006760889043410619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,7168,0.1120791170332167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,6144,0.11105600330564712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,768,0.007340444458855524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,768,0.00756444450881746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,8192,0.0477182235982683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,1536,0.009362666971153682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,7168,0.22964978218078613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,16384,0.039558221896489464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,2560,0.016728000508414376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,2048,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,128,0.0028648889727062653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,128,0.006934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,16384,0.028146667612923518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,6144,0.009393778112199571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,1536,0.02602133320437537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,1024,0.02661688956949446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,65536,0.06818399826685588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,3584,0.010068444742096795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,1024,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,7168,0.02816177739037408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,256,0.006090666684839461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,12288,0.06212355693181356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,64,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,2560,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,12288,0.024759110477235582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,12288,0.5351902114020454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,3584,0.012454222473833295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,6144,0.02014666630162133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,7168,0.014501333236694336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,10240,0.021888888544506494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,4096,0.007652444144090016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,10240,0.011744000017642975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,12288,0.03805955582194858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,1024,0.02735200027624766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,10240,0.012135110795497894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,65536,0.09303110837936401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,12288,0.030950221750471327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,5120,0.008908444808589088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,2560,0.007359999749395583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,4096,0.023824888798925612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,128,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,256,0.005594666633341048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,16384,0.046205331881841026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,7168,0.027463111612531874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,10240,0.017720000611411203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,128,0.01657777859105004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,512,0.003559111307064692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,128,0.002883555574549569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,7168,0.02734044525358412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,3072,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,6144,0.01720266706413693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,64,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,8192,0.061365332868364125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,3072,0.047396444612079196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,65536,0.05693333016501533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,768,0.015466666883892484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,512,0.00350222239891688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,128,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,7168,0.013351110948456658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,16384,0.1980266703499688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,1536,0.026520889666345384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,4096,0.04132444328731961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,128,0.0031555555760860443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,32,0.012456888953844706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,7168,0.009663110805882348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,512,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,3072,0.007307555940416124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,3072,0.018206222189797294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,4096,0.0199617776605818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,7168,0.007997333175606197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,4096,0.012823111481136747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,16384,0.02734755476315816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,10240,0.40027909808688694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,8192,0.04954399996333652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,3584,0.027247111002604168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,65536,0.051735109753078885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,5120,0.01959288948112064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,512,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,2048,0.009292444421185387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,1024,0.0144213338692983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,1024,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,4096,0.011066666907734342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,2560,0.007620444728268518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,12288,0.08637866708967422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,1024,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,768,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,128,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,6144,0.018888889087571036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,5120,0.05495288968086243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,16384,0.02141155633661482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,512,0.012080888781282636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,4096,0.133696887228224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,4096,0.12251199616326226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,1024,0.00737955586777793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,3072,0.014101333088344999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,6144,0.022612444228596155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,12288,0.017621333400408428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,64,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,256,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,2048,0.013439111411571503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,2560,0.007248889240953658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,768,0.004589333302444882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,16384,0.026399110754330952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,12288,0.05953866905636258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,2560,0.06002044677734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,3584,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,3584,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,8192,0.03295110993915134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,1024,0.0266266663869222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,4096,0.015232889188660515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,3072,0.011694221860832639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,1024,0.0038728887836138406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,768,0.016484444340070088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,8192,0.03223377797338698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,7168,0.009020444419648912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,1024,0.01584533353646596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,10240,0.15489333205752903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,128,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,7168,0.012171555724408893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,12288,0.01312622262371911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,256,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,4096,0.04118400149875217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,12288,0.04570133487383524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,3584,0.00903555585278405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,128,0.023915555742051866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,7168,0.027592889136738245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,7168,0.025640888346566096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,32,0.003177777760558658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,5120,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,16384,0.2548728783925374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,4096,0.013139555851618448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,7168,0.03569866551293267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,10240,0.015458666616015963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,2560,0.016199999385409884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,3584,0.06046311060587565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,128,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,64,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,10240,0.02409244411521488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,16384,0.015933333171738517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,1024,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,1024,0.007648000286685095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,768,0.07616977559195624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,3584,0.008861333131790161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,4096,0.029688000679016113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,12288,0.02907199992073907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,256,0.009992000129487779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,6144,0.013340444200568728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,768,0.03992444607946608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,12288,0.06093155675464206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,1024,0.014519999424616495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,6144,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,12288,0.02034311162100898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,16384,0.008966222405433655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,3072,0.05464799867735969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,7168,0.06683822472890218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,7168,0.02221422228548262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,12288,0.018576888574494254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,65536,0.06312621964348687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,2048,0.006612444503439798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,512,0.014717333846622042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,16384,0.02179466684659322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,65536,2.0462044609917536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,16384,0.030844443374209937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,7168,0.12197777960035537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,768,0.004568888909286923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,4096,0.005681777579916849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,6144,0.010339555641015371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,65536,4.180992974175347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,128,0.016927111479971144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,512,0.01038577738735411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,2560,0.008012444608741337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,2048,0.008424888882372115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,8192,0.010684444672531553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,64,0.006583999842405319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,768,0.01646133263905843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,2560,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,12288,0.024012444747818842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,2560,0.006903111106819577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,1024,0.007272889216740926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,128,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,16384,0.02642755541536543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,32,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,3072,0.03808000021510654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,768,0.0063386667105886675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,12288,0.0660346680217319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,32,0.003839111162556542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,2560,0.006992888947327931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,128,0.01145866678820716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,10240,0.012816888590653738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,5120,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,10240,0.0151582227812873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,128,0.011023110813564725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,65536,0.14018577999538845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,16384,0.23056088553534615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,512,0.009328888522254096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,8192,0.016891555653678048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,12288,0.16604088412390813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,7168,0.013269333375824822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,10240,0.026759111218982275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,1024,0.014760888285107084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,6144,0.06059555874930488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,65536,0.0595466693242391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,64,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,2048,0.022424888279702928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,65536,0.07804177867041694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,65536,1.1354328791300456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,2560,0.06773689058091906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,3584,0.016207110550668504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,16384,0.05474933319621616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,1536,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,65536,0.5406319830152724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,3584,0.01036088913679123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,65536,0.06668800115585327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,4096,0.010799999866220685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,3072,0.01847555571132236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,2048,0.011383111278216044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,7168,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,3584,0.009671111073758867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,3072,0.014131555954615274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,3072,0.016560888952679105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,1536,0.08061689138412476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,768,0.06606311268276639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,1536,0.004555555681387584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,5120,0.024840889705551997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,32,0.019856888386938307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,65536,0.19221156173282197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,16384,0.03412533468670315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,7168,0.011415999796655444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,1536,0.004892444445027245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,65536,0.08394844664467706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,2048,0.00703999979628457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,7168,0.025518221987618342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,1024,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,3584,0.1466568840874566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,7168,0.02198311189810435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,768,0.003896000070704354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,2560,0.019800000720553927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,8192,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,2048,0.006538666784763336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,3584,0.15369155671861437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,1536,0.022611555125978258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,16384,0.025514667232831318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,256,0.01909244391653273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,8192,0.018552889426549275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,8192,0.2339279916551378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,12288,0.020133333073721994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,2560,0.014578666951921252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,10240,0.04831377665201823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,768,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,512,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,256,0.006593777901596493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,10240,0.017755554782019723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,2048,0.012089778151777057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,7168,0.024354666471481323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,3584,0.01751377847459581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,256,0.015412444869677225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,16384,18.563700358072918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,256,0.012764444781674279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,32,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,5120,0.18977777163187662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,2048,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,128,0.0147386673423979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,8192,0.011208888557222156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,7168,0.010853333605660332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,12288,0.03204444381925795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,256,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,128,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,7168,0.019342222147517733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,5120,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,1024,0.006262222097979651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,512,0.004247111164861255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,128,0.016223111086421542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,65536,0.01498311095767551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,16384,0.021552888883484736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,3072,0.023445333043734234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,5120,0.009009777671760982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,2048,0.01203644441233741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,5120,0.056369774871402316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,128,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,5120,0.026877333720525105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,32,0.004571555389298333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,5120,0.012429333395428128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,32,0.0029031110720502008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,16384,0.024508444799317256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,2560,0.00755466686354743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,2048,0.012443555725945367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,7168,0.010207110808955299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,768,0.01629244453377194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,6144,0.009349333743254343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,256,0.0032231110251612137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,16384,0.1415404478708903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,2560,0.014211555322011312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,64,0.0028968888024489083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,128,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,6144,0.02490933405028449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,12288,0.01862133377128177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,2048,0.00720355576939053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,512,0.010086221827401055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,3072,0.015060444672902426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,12288,0.018473777506086562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,1536,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,2048,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,6144,0.03087110983000861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,768,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,8192,0.12771199809180364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,1536,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,8192,0.014078223043017916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,8192,0.3624586794111464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,65536,0.020008888509538438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,2048,0.08707644542058308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,3584,0.0165964447789722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,128,0.003903999924659729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,7168,0.04570844438340929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,12288,0.11263288391960992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,512,0.007234666910436418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,64,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,10240,0.010402667025725046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,768,0.021898667017618816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,3584,0.01350933313369751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,65536,0.29997777938842773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,12288,0.13902310530344644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,6144,6.982045491536458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,6144,0.04618933465745714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,6144,0.03231466809908549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,64,0.004911111046870549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,6144,0.036158220635520086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,768,0.025049777494536504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,1536,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,10240,0.02198666665289137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,64,0.0069448889957533936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,1024,0.01718755563100179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,512,0.00702400008837382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,64,0.00350577798154619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,4096,0.16444888379838732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,16384,0.02223288847340478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,3584,0.015044444137149386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,12288,0.009559111462699043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,128,0.0031146665828095544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,4096,0.024989333417680528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,2560,0.04229688975546095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,3584,0.011021333436171213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,1536,0.02788355615403917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,3072,0.013069333301650153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,768,0.016912000046836007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,64,0.006238222122192383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,6144,0.0090844440791342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,3584,0.3356417814890544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,64,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,6144,0.47909869088066953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,10240,0.02200177808602651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,10240,0.1193288829591539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,10240,0.045272890064451426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,512,0.006325333482689328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,10240,0.02574133376280467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,32,0.007963555554548899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,32,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,10240,0.025176889366573755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,4096,0.007618666523032718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,256,0.009662222531106737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,2048,0.008288000192907121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,1024,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,3584,0.013432888521088494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,64,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,8192,0.18312266137864855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,3584,0.028976887464523315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,128,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,1024,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,6144,0.019955555597941082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,256,0.0033048888047536216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,16384,0.04790666699409485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,8192,0.03010222315788269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,7168,0.02020088831583659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,7168,0.10580977466371323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,3072,0.007677333222495184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,512,0.01853155593077342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,64,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,8192,0.022597332795461018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,7168,0.019142222073343065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,3584,0.007096889118353526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,2048,0.006699555450015598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,1024,0.014097778333557976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,1024,0.01793866687350803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,512,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,12288,0.16061688794030085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,32,0.0052693333062860705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,7168,0.024315555890401203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,6144,0.02201066745652093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,2048,0.11073511176639134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,16384,0.03336888882848952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,768,0.015809777710172866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,4096,0.00795288880666097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,256,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,4096,0.008373333348168267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,10240,0.02640977833006117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,8192,0.0070186663005087115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,3072,0.007293333609898885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,5120,0.011408000356621213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,2560,0.033737778663635254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,1024,0.03948533203866746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,32,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,5120,0.02741066614786784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,6144,0.024338665935728285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,6144,0.01108088923825158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,2560,0.012453333371215396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,1024,0.008940444224410588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,12288,0.026263111167483862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,256,0.008018666671382057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,3072,0.00794044468137953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,64,0.003212444484233856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,4096,0.011228444675604502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,3072,0.008292444050312042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,2560,0.0069919998447100324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,65536,0.05410222212473551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,2560,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,65536,1.914105733235677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,1536,0.015506666567590503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,7168,0.14263644483354357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,4096,0.01717866626050737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,128,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,1024,0.025855110751258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,768,0.0038444445365005066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,2048,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,16384,0.027477333943049114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,16384,0.2325057718488905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,4096,0.011720889144473605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,12288,0.017839999662505258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,7168,0.01460444430510203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,65536,0.07720888985527886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,32,0.0042133331298828125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,2560,0.03164533442921109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,12288,0.03781244489881728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,8192,0.02126933303144243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,2560,0.07912444406085543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,2560,0.01888444523016612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,256,0.006301333506902059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,6144,0.010017777482668558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,12288,0.0131217779384719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,512,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,3584,0.008099555969238281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,3072,0.023006222314304773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,768,0.007848888635635376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,256,0.005207111024194294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,5120,0.017364444004164804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,12288,0.03569422165552775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,768,0.009224889179070791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,1536,0.01891377733813392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,2560,0.04820355441835192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,6144,0.010031110710567897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,10240,0.019223110543357003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,6144,0.007456888755162557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,768,0.018627555833922494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,128,0.013811555173661975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,2560,0.007359999749395583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,12288,0.03225422236654494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,4096,0.01759466694460975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,2048,0.006586666736337874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,12288,0.059270222981770836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,12288,0.05384888913896349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,256,0.004264889078007804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,128,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,12288,0.03936355643802219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,1536,0.015190222197108798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,8192,0.01038933296998342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,128,0.014146667387750415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,2560,0.008026666939258575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,1536,0.023909333679411147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,256,0.006636444479227066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,1024,0.003624000069167879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,65536,0.08611377742555405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,65536,2.799594667222765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,512,0.008806222014957005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,6144,0.023959999283154804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,12288,0.03585422370168898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,3584,0.014582221706708273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,10240,0.012169777519173093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,1536,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,32,0.00794311116139094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,65536,0.11608978112538655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,4096,0.019919999771647982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,3584,0.7138951089647082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,3072,0.028400000598695543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,768,0.03289511137538486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,3584,0.013450666434235044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,2048,0.0069955554273393415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,64,0.0032702222880389956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,8192,0.036074668169021606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,65536,0.05680533250172933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,32,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,2048,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,768,0.006061333335108227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,64,0.0028560000161329904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,12288,0.016200888488027785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,1024,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,6144,0.051386667622460254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,10240,3.041788525051541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,1536,0.024718221690919664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,512,0.02269511090384589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,512,0.01629599928855896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,512,0.049535112248526685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,128,0.0032426667296224167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,16384,0.02223377757602268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,3072,0.00797333319981893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,6144,0.021887999441888597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,2048,0.012859555582205454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,12288,3.2802408006456165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,1024,0.0166311115026474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,768,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,8192,0.027667555544111464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,3584,0.026567111412684124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,3584,0.02369955513212416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,3584,0.014176888598336114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,512,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,3584,0.007692444655630324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,512,0.014488000008794995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,3072,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,4096,0.027994665834638808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,3584,0.016750221451123554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,6144,0.018952000472280715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,128,0.016843555702103507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,512,0.0038737778862317405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,512,0.030968000491460163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,65536,0.0755760007434421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,3584,0.024412444896168176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,1024,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,256,0.015504888362354703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,10240,0.06589066982269287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,256,0.03155111273129781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,5120,0.23458578851487902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,3584,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,6144,0.047397332059012524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,16384,0.22199643982781303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,256,0.003144888828198115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,10240,0.21112710899776885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,8192,0.012319111161761813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,64,0.009361777868535783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,1536,0.016553777787420485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,8192,0.021659556362364028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,768,0.007919111185603673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,32,0.0028479999552170434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,32,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,6144,0.12638933128780788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,16384,0.3445777893066406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,16384,0.03901244534386529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,1536,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,12288,0.09908088710572983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,6144,0.012116444607575735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,768,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,256,0.014847111370828418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,2560,0.016234666109085083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,10240,0.009676444033781687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,3072,0.022778666681713525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,64,0.006294222341643439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,5120,0.017208000024159748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,2048,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,65536,0.03605511122279697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,65536,0.06681866778267755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,512,0.014090667168299357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,16384,0.08544266886181301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,4096,0.1810284455617269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,16384,0.0583546691470676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,1536,0.008654221892356873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,256,0.01793688866827223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,768,0.01687111126052009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,512,0.015247111519177755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,6144,0.010063110954231685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,8192,0.06971822182337443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,1536,0.004547555413511065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,65536,0.06439466608895196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,64,0.009296000003814697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,2048,0.010696888797812991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,2048,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,5120,0.02536444365978241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,65536,0.0709457794825236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,3072,0.007252444823582967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,65536,0.7332737710740832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,64,0.0032640000184377036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,16384,0.039084444443384804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,12288,0.04598133431540596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,768,0.015082667271296183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,768,0.017533333765135873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,10240,0.022974222898483276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,8192,0.033942222595214844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,10240,0.07724799712498982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,16384,0.06453777684105767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,6144,0.013418667018413544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,2048,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,1536,0.11834755208757188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,5120,0.010279110736317104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,4096,0.011109333071443768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,512,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,1536,0.02349866595533159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,16384,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,65536,0.009273777405420939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,128,0.014817777607176038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,16384,0.04054666558901469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,3072,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,128,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,16384,0.04948000113169352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,16384,0.013570666313171387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,5120,0.027298667364650305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,5120,0.015600888265503777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,4096,0.055323554409874805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,1536,0.019409777389632333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,512,0.013376889129479727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,12288,0.03258400162061056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,512,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,2048,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,16384,3.2354479895697703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,12288,0.017824000782436795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,128,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,3584,0.009660444325870937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,512,0.01348622226052814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,8192,0.1365262269973755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,16384,0.03304533494843377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,12288,0.2871493233574761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,256,0.011597333682907952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,7168,0.01889866590499878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,6144,0.39493510458204484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,8192,0.0654675563176473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,65536,0.6687431335449219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,12288,0.2733626630571154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,8192,0.02808444367514716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,768,0.016369778248998854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,12288,0.0487955543729994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,2048,0.019682667321628995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,2560,0.015105777316623263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,5120,0.02757155564096239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,1536,0.007689778175618913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,2048,0.021525333325068157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,512,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,64,0.006216889040337668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,1024,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,65536,0.13978311750623915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,1536,0.010455999937322406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,12288,0.0206986665725708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,2560,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,768,0.01995111174053616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,16384,0.031179553932613794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,768,0.017242666747834947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,3072,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,256,0.003236444460021125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,65536,0.08534133434295654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,5120,0.00942133367061615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,65536,0.14192533493041992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,16384,0.08735911051432292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,12288,0.009713778065310584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,3072,0.007680888805124495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,65536,0.11902488602532281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,10240,0.011420444481902652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,1536,0.014422222971916199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,6144,0.022775999373859827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,5120,0.012124444047609964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,10240,0.01591199967596266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,65536,0.07880622148513794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,7168,0.014428445034556918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,768,0.045646223757002086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,3072,0.013470222552617392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,256,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,1024,0.005631999837027655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,128,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,12288,0.03233155608177185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,12288,0.018265777164035372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,1024,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,256,0.01125955581665039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,4096,0.007947555846638149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,16384,0.008938666847017076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,4096,0.03260533346070184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,65536,0.07484089003668891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,5120,0.023147554861174688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,512,0.015519110692871941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,3072,0.0429075558980306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,768,0.007727111379305522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,10240,0.023726221587922838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,128,0.006073777874310811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,2560,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,32,0.0028702221396896574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,5120,0.1980311075846354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,8192,0.01441511180665758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,128,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,8192,0.036658667855792575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,3584,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,3584,0.015133332875039844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,5120,0.4571173455980089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,64,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,256,0.007925333248244392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,32,0.008347555167145198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,6144,0.01386755539311303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,16384,0.021330666210916307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,768,0.03165333469708761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,6144,0.01262399968173769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,3584,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,1536,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,512,0.012146666646003723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,3072,0.022325333621766832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,3072,0.23745867941114637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,3072,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,32,0.054132444991005786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,768,0.016892444756295946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,12288,0.04765689041879442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,128,0.0031439999325407874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,4096,0.01882577770286136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,2048,0.025425778494940862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,65536,0.04983466532495287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,512,0.025412445267041523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,32,0.00350222239891688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,2560,0.014807111687130399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,512,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,1024,0.014185777968830533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,1536,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,512,0.0034942221310403612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,10240,0.025993777645958796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,6144,0.12401866912841797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,16384,0.015436444017622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,65536,0.3151902092827691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,4096,0.009145778086450364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,7168,0.20578666528066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,4096,0.01202488856183158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,2048,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,5120,0.01571555601225959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,3072,0.017464000317785475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,256,0.014831110835075378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,8192,0.02397866712676154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,12288,0.032391998502943255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,7168,0.012100444071822695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,16384,0.2181839942932129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,8192,0.012054222325483957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,5120,0.01819022165404426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,768,0.01590222285853492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,256,0.005568000177542369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,32,0.002817777709828483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,12288,0.1706542174021403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,256,0.012991999586423239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,6144,0.017228444417317707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,2560,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,32,0.0024951110697454875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,2560,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,5120,0.012774222426944308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,768,0.01681155628628201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,16384,0.03668266534805298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,1536,0.010660444696744284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,2048,0.007982222570313348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,8192,0.007258666886223688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,12288,0.03254488772816128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,8192,0.029331554969151814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,12288,0.0166311115026474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,12288,0.129077328575982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,1024,0.025359110699759588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,256,0.016487111647923786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,6144,0.022621333599090576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,256,0.004848888764778773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,5120,0.01035199976629681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,256,0.007646222081449296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,16384,0.022121777137120564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,256,0.024001777172088623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,2048,0.09942489200168186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,512,0.0035751110149754416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,512,0.006281777802440856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,65536,0.09272533655166626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,1536,0.03805333375930786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,5120,0.008623110751310984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,12288,0.019676445258988273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,128,0.01827022267712487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,256,0.009705777797434065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,8192,0.05240266521771749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,16384,0.05036355720625984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,128,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,8192,0.010680889089902243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,2560,0.03296800123320685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,128,0.004945777770545748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,16384,0.2140737771987915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,256,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,2048,0.010994666980372535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,3584,0.053704000181621976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,5120,0.02126844392882453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,8192,0.013532444834709167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,10240,0.012049777640236748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,12288,0.007263999846246507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,6144,0.01887822151184082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,16384,0.03637689020898607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,256,0.020241777102152508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,1536,0.017581333716710407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,3584,0.007576888634098901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,128,0.010669333239396414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,7168,0.021074665917290583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,256,0.00619911112719112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,7168,0.01886133352915446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,4096,0.01606755620903439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,3584,0.09481155872344971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,16384,0.14431377251942953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,128,0.016522667474216886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,768,0.006032889087994893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,8192,0.021993777818149988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,8192,0.4189937909444173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,4096,0.03322488731808133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,8192,0.012081777883900536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,10240,0.007994666695594788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,3072,0.023011555274327595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,1536,0.015241778559154935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,768,0.01575911045074463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,3584,0.009916444619496664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,10240,0.04093777802255418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,6144,0.0899111098713345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,1536,0.004592888885074192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,10240,0.12434666686587864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,65536,0.06597066587871976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,3072,0.008642666869693333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,10240,0.03471111257870992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,16384,0.048955556419160634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,7168,0.17113510767618814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,8192,0.045556445916493736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,2560,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,12288,0.29890399509006077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,1024,0.015519110692871941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,65536,1.187321768866645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,65536,0.21075199709998238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,5120,0.0287715560860104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,2560,0.00962400022480223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,10240,0.016369778248998854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,3584,0.014055111342006259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,3584,0.020418667131000094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,12288,0.03695111142264472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,5120,0.017044444878896076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,4096,0.09176444345050389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,16384,0.03661777906947666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,3072,0.08243822389178805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,128,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,65536,0.37660800086127383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,1024,0.009684444301658207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,256,0.006945778098371293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,1024,0.007979555262459649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,128,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,512,0.003847111016511917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,1024,0.0038693332009845306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,256,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,1536,0.018836443622907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,2560,0.007183999651008182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,8192,0.015104000767072042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,4096,0.008280889027648503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,1024,0.008172444171375698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,6144,0.011941333611806234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,256,0.00370755543311437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,8192,0.02253866692384084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,1536,0.00685066646999783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,7168,0.01961599952644772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,6144,0.06578133503595988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,10240,0.23958044581943086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,4096,0.06204622321658664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,1024,0.007359111474619971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,1024,0.014244443840450711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,768,0.06268977456622653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,4096,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,16384,0.12813956207699245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,64,0.0032311110860771606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,3072,0.00794577764140235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,12288,0.040919999281565346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,128,0.023715555667877197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,10240,0.038592000802357994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,3584,0.008360000120268928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,8192,0.01943466729587979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,2048,0.054233776198493115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,6144,0.03449155555831061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,10240,0.018380444910791185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,2048,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,768,0.010703110860453712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,512,0.005428444594144821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,16384,0.05455466773774889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,4096,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,512,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,4096,0.01687022215790219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,2560,0.0069955554273393415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,10240,0.0325617790222168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,1024,0.06122399701012505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,256,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,512,0.010027555955780877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,512,0.013461333182122974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,8192,0.02899200055334303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,2560,0.0823226637310452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,10240,0.5182106759813097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,10240,0.06480889187918769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,64,0.009317333499590555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,32,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,4096,0.007298666569921706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,10240,0.018578666779730055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,16384,0.028439998626708984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,4096,0.011028444601429833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,256,0.0069164443347189165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,16384,0.03179377648565505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,2048,0.005561777700980504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,4096,0.017889777819315594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,256,0.010453333457310995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,128,0.013757333159446716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,4096,0.18005777729882133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,4096,0.06226488616731432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,256,0.016526222229003906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,1536,0.01275199982855055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,32,0.006896889044178857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,128,0.0035137778355015647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,3584,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,1536,0.013085333009560903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,1536,0.006706666615274217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,2560,0.01237066669596566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,2560,0.02228533393806881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,5120,0.01199288914601008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,512,0.0035342222286595237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,10240,1.8050346374511719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,512,0.02509333358870612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,512,0.006311999840868845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,10240,0.0107360002067354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,3584,0.012963555753231049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,65536,0.4203546577029758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,5120,0.00903733323017756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,10240,0.36417243215772843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,12288,0.038690666357676186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,12288,0.014097778333557976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,1024,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,2048,0.01424533294306861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,512,0.01071200023094813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,128,0.01277599980433782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,16384,0.013157332936922709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,768,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,4096,0.0675742228825887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,12288,0.00831822223133511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,12288,0.013232000172138214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,64,0.004202666795916027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,512,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,10240,0.03860355416933695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,1536,0.0069146665434042616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,6144,0.009704888694816166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,4096,0.013467555244763693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,3072,0.09285689062542385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,10240,0.02196088929971059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,1024,0.014117333624098035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,2560,0.077766219774882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,5120,0.025968000292778015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,8192,0.045702222320768565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,32,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,768,0.006337777607970768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,2048,0.014939554863505893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,1024,0.007829333345095316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,5120,0.025616000096003216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,2048,0.012433778080675336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,1024,0.007696000238259633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,7168,0.03201155530081855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,1536,0.016189333465364244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,5120,0.011121778024567498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,3072,0.0064284445510970215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,32,0.003556444413132138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,16384,0.0472871098253462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,7168,0.020987555384635925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,1536,0.09992088874181111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,6144,0.00904444439543618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,3584,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,1536,0.024608888559871252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,6144,0.024130667249361675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,12288,0.04298311140802172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,8192,0.024922667278183833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,5120,0.008864888714419471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,2560,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,8192,0.013812444276279874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,3072,0.006950221955776215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,512,0.005203555441564984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,768,0.008643555144468943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,8192,0.033448888195885554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,1536,0.07265688975652058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,16384,0.044806222120920815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,1024,0.014483556151390076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,256,0.016509332590632968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,6144,0.0075697774688402815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,6144,0.01311911145846049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,3584,0.010582221878899468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,768,0.0057839999596277876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,8192,0.00869511150651508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,65536,0.09551111194822524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,8192,0.020958221620983545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,1024,0.007979555262459649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,10240,0.02827199962404039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,128,0.0032151111712058387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,2560,0.011780444118711682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,32,0.011807111402352652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,256,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,64,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,8192,0.01956266661485036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,3072,0.018999111321237352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,3072,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,2560,0.2525537808736165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,8192,0.010633777413103314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,2560,0.0073724447025193115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,3072,0.022971555590629578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,2560,0.03388266762097677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,3072,0.02440888848569658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,2048,0.01036177741156684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,2048,0.014408000641398959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,128,0.0035297779573334586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,12288,0.05870311127768623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,256,0.04330577784114414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,512,0.00535377777285046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,8192,0.07457066906823052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,5120,0.017921777235137094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,1024,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,6144,0.028059555424584284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,5120,0.05384711093372769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,2560,0.016833777228991192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,16384,0.03739022215207418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,2048,0.014621333943472968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,64,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,1536,0.014162666267818876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,6144,0.018555555078718398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,64,0.002716444432735443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,256,0.006586666736337874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,65536,0.04998844530847338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,1536,0.01683022247420417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,6144,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,3072,0.023042667243215773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,12288,0.049638221661249794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,3584,0.2028986612955729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,8192,0.0247999992635515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,5120,0.033936887979507446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,1536,0.015545777148670621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,2560,0.00758844448460473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,2560,0.03605422377586365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,16384,1.4800390667385523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,6144,0.018726223044925265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,6144,0.02681866619322035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,2048,0.006592000110281839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,32,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,64,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,64,0.0035093331502543557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,8192,0.029657777812745836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,3072,0.02753066685464647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,64,0.006615110983451207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,8192,0.015460444821251763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,4096,0.08431111441718207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,12288,0.09679288996590509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,12288,0.020350222786267597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,10240,0.021623111433453027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,256,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,1024,0.006158222340875202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,3584,0.0885724425315857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,5120,0.06804888778262667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,6144,0.02104444470670488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,4096,0.07388710975646973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,1536,0.03474488854408264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,7168,0.014660444524553088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,3072,0.00981955561372969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,768,0.006044444524579578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,65536,0.07559555768966675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,512,0.003584889074166616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,7168,0.028472890456517536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,256,0.017176000608338248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,65536,0.09726044204499985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,768,0.0038951109680864546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,3584,0.018757333358128864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,2048,0.07239200009240045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,10240,0.016729777057965595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,3072,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,6144,0.02455288834042019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,1536,0.006745777610275481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,768,0.005772444523043103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,16384,0.020403555697864957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,2560,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,12288,0.02751200066672431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,6144,0.006575999988449945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,3584,0.00792622235086229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,8192,0.015562666787041558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,4096,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,6144,0.14892621835072836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,12288,0.0488302210966746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,3072,0.012824888858530255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,1536,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,512,0.012068444656001197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,3584,0.01570222278436025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,768,0.004212444441186057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,16384,0.03631466627120972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,65536,0.09851555691825019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,256,0.006839111033413146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,3584,0.018979556030697294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,512,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,1024,0.017433777451515198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,256,0.0035173334181308746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,128,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,64,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,768,0.014583999911944071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,16384,0.016558221644825406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,128,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,5120,0.009963555468453301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,2560,0.007309333317809635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,2560,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,256,0.022022222479184467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,32,0.0031519999934567344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,32,0.010346666806273991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,10240,0.15041422843933105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,768,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,6144,0.37763200865851504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,768,0.022011554903454248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,32,0.013767999907334646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,4096,0.19158489174313018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,2560,0.011345778074529437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,1536,0.013167999684810638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,1024,0.01906399925549825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,4096,0.029300444655948218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,7168,0.01482488877243466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,8192,0.03978577918476529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,12288,0.03195644418398539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,3072,0.011191111471917896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,64,0.0031902222997612427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,7168,0.014197332991494073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,2048,0.05585777759552002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,256,0.009840889109505547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,7168,0.012621333201726278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,3072,0.011306666665607028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,2560,0.007655111451943715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,128,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,256,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,512,0.0035093331502543557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,65536,0.08644355667961968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,512,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,512,0.00591822216908137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,6144,0.024137778414620295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,128,0.005215999980767568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,256,0.005548444473081165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,1024,0.021700445148679946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,768,0.004081777814361784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,4096,0.01239999963177575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,10240,0.14005155033535427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,2560,0.016559110747443307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,12288,0.05964799722035726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,2560,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,10240,0.011762667033407422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,768,0.007262222468852997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,64,0.016199999385409884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,768,0.006285333385070165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,128,0.014418666561444601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,2560,0.017107556263605755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,10240,0.03650755683581034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,6144,0.012796444197495779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,2560,0.016749334004190233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,7168,0.021848888860808477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,4096,0.027074666486846075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,768,0.022247110803922016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,10240,0.03699644406636556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,2560,0.007387555307812161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,512,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,10240,0.5665609041849772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,512,0.007595555649863348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,7168,0.02223644488387638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,16384,0.03177333209249709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,8192,0.018585777944988675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,2560,0.01949155496226417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,10240,0.012128888732857175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,512,0.0032115555885765287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,4096,0.038754665189319186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,12288,0.01593155496650272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,256,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,5120,0.007684444387753804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,1536,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,64,0.0028773333049482773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,256,0.015438222222858004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,1024,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,5120,0.008625778059164682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,3072,0.007624000310897827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,3072,0.010615111225181155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,65536,0.5017111036512587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,1024,0.009645333720578088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,3072,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,5120,0.012794666820102267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,8192,0.07877333296669854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,3584,0.017110221915774874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,1024,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,12288,0.19279821713765463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,16384,0.015551111764378019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,256,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,12288,0.013167111410035027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,65536,0.009002666506502364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,128,0.015138667490747241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,7168,0.061965333090888135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,512,0.011360888679822287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,64,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,2560,0.05704800287882487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,10240,0.011886222494973077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,3072,0.007255111303594377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,16384,0.01584533353646596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,6144,0.016349333855840895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,128,0.00526577772365676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,3072,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,12288,0.0773164431254069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,512,0.016548444827397663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,4096,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,256,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,1536,0.008630221916569604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,6144,0.08226755592558119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,3584,0.00796088907453749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,1024,0.004559111264016894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,12288,0.03068800104988946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,3584,0.009099555512269339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,128,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,128,0.015138667490747241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,3072,0.00777422222826216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,4096,0.013816888961527081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,4096,0.016885333591037326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,8192,0.00828888929552502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,12288,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,1024,0.020264888803164165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,5120,0.008960000342792934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,2560,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,256,0.00833155545923445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,3072,0.015251555376582675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,7168,0.03167288833194309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,256,0.015473778049151102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,6144,0.013497778111033969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,3584,0.007465778125656976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,3584,0.026791999737421673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,1536,0.009530666801664565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,128,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,12288,0.041856888267729014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,1024,0.012700444294346703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,512,0.03188177943229675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,3072,0.013442666994200813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,256,0.007033777733643849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,1024,0.012096000214417776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,65536,0.0586400032043457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,768,0.025255110528733995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,3072,0.0991671085357666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,3072,0.04944088723924425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,12288,0.027464888162083093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,2048,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,4096,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,512,0.03394577900568644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,4096,0.010471999645233154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,6144,0.13224355379740396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,2560,0.005738666488064661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,1536,0.00870222184393141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,256,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,7168,0.012415111064910889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,10240,0.03289511137538486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,256,0.005397333453098933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,3072,0.011024889018800525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,3584,0.016794666647911072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,256,0.0052355556852287715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,1536,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,128,0.0032311110860771606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,2048,0.01830666595035129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,4096,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,5120,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,7168,0.015426667200194465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,65536,0.32644889089796275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,32,0.02067466742462582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,7168,0.014694222145610385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,16384,0.05999910831451416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,32,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,6144,0.05482222305403816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,1536,0.014198222094111972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,1536,0.006288888967699475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,16384,0.022109333011839125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,7168,0.05121866530842251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,2560,0.007272889216740926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,4096,0.008068444828192392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,2560,0.03087377879354689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,2048,0.009218666288587783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,5120,0.009716444545321995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,4096,0.008258666429254744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,128,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,2560,0.02919644448492262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,2560,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,12288,0.009761778016885122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,3072,0.17027821805742052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,768,0.003818666769398583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,10240,0.017024000485738117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,12288,0.0311119986904992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,3072,0.01035911093155543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,10240,0.03829955392413669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,2048,0.06131911277770996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,4096,0.030538668235143025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,2048,0.03650044401486715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,128,0.01480977733929952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,65536,0.08322933647367689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,4096,0.02230577833122677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,8192,0.008575110799736446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,8192,0.16990666919284395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,2048,0.006835555450783835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,2048,0.015035554766654968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,2560,0.006319111006127463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,4096,0.019243554936514962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,3584,0.008033778104517195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,128,0.020921778347757127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,5120,0.10958666933907403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,2048,0.006969778074158563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,5120,0.015677332878112793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,10240,0.4598871337042914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,3584,0.1798079941007826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,4096,0.16580711470709905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,7168,0.01292799992693795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,128,0.0031511110977994073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,1536,0.016206221448050607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,8192,0.7657901975843641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,4096,0.16868444283803305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,16384,0.04587022132343716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,1024,0.019278221660190158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,7168,0.03163377775086297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,7168,0.1537919971677992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,7168,0.010404444403118556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,2048,0.017138666576809354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,3584,0.09270399808883667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,3072,0.016171556380059984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,512,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,8192,0.2970613373650445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,768,0.016519111063745286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,1536,0.014779556128713818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,16384,0.16045155790117052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,3584,0.014418666561444601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,6144,0.009633777870072259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,65536,0.08667021989822388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,128,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,768,0.003569777641031477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,7168,0.011872000164455838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,256,0.0052426668504873914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,7168,0.010614222122563256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,2048,0.00758755538198683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,512,0.005210666606823604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,12288,0.03687644335958693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,256,0.015407999356587728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,128,0.005647111270162795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,768,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,10240,0.1649493376413981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,768,0.005371555685997009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,2560,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,6144,0.010832000109884473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,12288,0.03009777930047777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,3584,0.008175111479229396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,6144,0.06075111362669203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,16384,0.5559964179992676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,8192,0.0422017772992452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,5120,0.022708444131745234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,768,0.011191999746693505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,256,0.007308444215191736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,8192,2.313430150349935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,1024,0.005921777751710679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,10240,0.012795555922720166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,32,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,65536,0.01551555593808492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,2048,0.0155395550860299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,2560,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,1024,0.0057760001056724125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,6144,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,10240,0.022327999273935955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,512,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,2048,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,1024,0.026132444540659588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,1536,0.008663111262851292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,16384,0.04635733366012573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,8192,0.02424977719783783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,2048,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,768,0.03718577822049459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,1536,0.006028444402747684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,3584,0.011064888702498542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,3072,0.010946667028797997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,768,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,10240,0.012746666868527731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,3584,0.015133332875039844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,65536,0.07272266679339938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,3584,0.014147554834683737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,768,0.014480888843536377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,1024,0.012469333079126147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,5120,0.008336000144481659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,5120,0.011174221833546957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,256,0.0032213332338465583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,2048,0.005251555393139522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,12288,0.019517333971129525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,12288,0.01754755609565311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,1024,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,768,0.0038844446341196695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,2048,0.05522400140762329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,65536,0.020398222737842135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,65536,0.23878399531046549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,768,0.006994666324721442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,65536,0.057075553470187716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,8192,0.019779556327395968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,3584,0.0210515558719635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,7168,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,65536,0.10094577736324734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,32,0.0031351111829280853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,2048,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,2560,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,3072,0.009393778112199571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,6144,0.008640000389681922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,1536,0.0063733334342638654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,4096,0.060942219363318555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,1024,0.007791111038790808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,64,0.0028106667515304354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,2560,0.0401760008600023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,12288,0.012936000194814471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,12288,0.04179911149872673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,2560,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,16384,0.36670488781399196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,1024,0.00832711077398724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,2048,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,512,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,128,0.013560000393125745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,256,0.011708444191349877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,1536,0.01683466633160909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,32,0.0028133332315418455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,2560,0.013372444444232516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,3072,0.02904355525970459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,1024,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,7168,0.051947553952534996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,2560,0.006989333364698622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,3584,0.010724444356229572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,1536,0.004577777865860197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,65536,0.05891199906667074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,128,0.0032266666077905228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,256,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,3584,0.015326221783955893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,16384,0.09355733129713271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,2048,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,4096,0.0085235552655326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,3072,0.00866844422287411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,768,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,2048,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,128,0.011725333001878528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,65536,0.09382666481865777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,256,0.006321777900060018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,256,0.02439555525779724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,512,0.0035084444615576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,10240,0.007637333538797166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,12288,0.028166221247778997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,10240,0.0321048895517985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,5120,0.008970666262838576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,2048,0.015111999379263984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,16384,0.2949368953704834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,12288,0.034549332327312894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,10240,0.30282844437493217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,7168,0.05865866608089871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,1024,0.0402124457889133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,3584,0.04146577914555868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,65536,0.08988889058430989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,10240,0.015582222077581616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,1536,0.010484444598356882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,8192,0.023392000132136877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,128,0.037800888220469155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,8192,0.03412977854410807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,3584,0.023431999815834895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,5120,0.008651555412345463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,7168,0.12451199690500896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,64,0.003917333152559069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,256,0.006699555450015598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,6144,0.033414221472210355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,2560,0.016992888516849942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,2560,0.006956444846259222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,512,0.006056889063782162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,7168,0.012345777617560493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,256,0.008049777812427944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,256,0.0031475555151700974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,1536,0.02159022291501363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,12288,0.03163733416133457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,8192,0.041069332096311785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,2560,0.11977956030103896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,3584,0.0075902218619982404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,1536,0.01906844476858775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,1536,0.008658666577604082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,1024,0.004018666843573253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,256,0.0030168888883458245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,64,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,3584,0.01386666629049513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,12288,0.19759821891784668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,3072,0.09173866775300767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,6144,0.010117332968446944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,512,0.00599644457300504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,5120,0.011824888487656912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,2560,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,32,0.005321777943107817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,65536,0.11798755327860515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,2048,0.013125333521101209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,1536,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,128,0.003573333223660787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,4096,0.014727110664049784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,12288,0.28491200341118705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,1024,0.013753777576817406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,256,0.008467555873923832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,512,0.00868000007337994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,3584,0.038557334078682795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,2048,0.008044444852405125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,12288,0.025594666600227356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,3072,0.02291999922858344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,64,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,4096,0.007976888782448238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,65536,0.1185493336783515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,1536,0.061499555905659996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,2560,0.009478222164842818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,7168,0.020512888828913372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,64,0.011352889239788055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,5120,0.009008000294367472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,1024,0.016563556260532804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,2048,0.11010133557849461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,512,0.007000889215204451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,8192,0.017512889371977914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,5120,0.013169777890046438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,3584,0.008630221916569604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,1536,0.02479466630352868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,4096,0.02277066641383701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,6144,0.00656088896923595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,12288,0.17849955293867323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,2560,0.00693866651919153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,2560,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,3072,0.12312889099121094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,768,0.019588443968031142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,1024,0.014110222458839417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,12288,0.017997332745128207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,2560,0.023178666830062866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,12288,0.020265777905782063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,256,0.0035582222044467926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,4096,0.021366222037209406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,4096,0.014969777729776172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,10240,0.06435288985570271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,3072,0.030753778086768255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,2560,0.007064888874689738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,1536,0.006575999988449945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,3072,0.015846222639083862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,512,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,64,0.0028124445428450904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,10240,0.20411377482944063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,10240,0.030408889055252075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,128,0.016513778103722464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,4096,0.013043555948469373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,2048,0.027941332923041448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,3072,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,1536,0.020632000433074105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,6144,0.4732506540086534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,3072,0.1524533298280504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,65536,0.10913511117299397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,8192,0.021959111094474792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,2048,0.017300445172521803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,6144,0.02978133492999607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,768,0.01000088867213991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,7168,0.023165333602163527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,65536,0.08693510956234402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,768,0.01579466627703773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,32,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,512,0.025256888733969793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,8192,0.013244444297419654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,3072,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,5120,0.026962666047943964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,2048,1.6166790856255426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,65536,0.11960444185468887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,2048,0.019482667247454327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,2048,0.016390222642156813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,6144,0.06701333655251397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,10240,0.03569511241383023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,65536,0.19013333320617676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,512,0.003912000192536248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,1024,0.07126666439904107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,7168,0.027335110637876723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,5120,0.017688888642523024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,2048,0.005498666730191972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,1536,0.00611200018061532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,1024,0.012982221941153208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,8192,0.03614933292071024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,2048,0.01444355481200748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,768,0.00563466673096021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,32,0.005636444522274866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,12288,0.04622755448023478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,2560,0.0222382214334276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,65536,0.11392266220516628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,6144,0.02566577825281355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,10240,0.04272355635960897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,12288,0.012074666718641916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,7168,0.015799111790127225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,8192,0.03163111209869385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,65536,0.10291110806994969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,2560,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,32,0.0028115556471877625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,2560,0.013070222404268054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,12288,0.01843377782238854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,64,0.002814222127199173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,3072,0.42640267478095156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,4096,0.01240799989965227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,4096,0.022073777185546026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,6144,0.007600888609886169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,768,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,16384,0.04067466656366984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,1024,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,65536,0.05537866552670797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,2560,0.017186666528383892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,1536,0.015878222054905362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,2560,0.013062222136391534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,3584,0.019267555740144517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,5120,0.008346666892369589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,512,0.00527022240890397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,12288,0.013458666702111563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,512,0.004906666775544484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,8192,0.016883555385801528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,128,0.003220444545149803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,6144,0.009672889278994666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,64,0.018800000349680584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,64,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,10240,0.01567911108334859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,5120,0.009397333694828881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,6144,0.012536889149083031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,3584,0.012772444221708508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,5120,0.00828888929552502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,4096,0.02115733259254032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,4096,0.012494222157531314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,2560,0.007335999773608313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,1536,0.005672889037264719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,512,0.003485333174467087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,768,0.02576444380813175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,4096,0.052609778112835355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,256,0.005940444353553984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,1024,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,64,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,6144,0.031014220582114324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,256,0.015182221929232279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,3584,0.02360800074206458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,512,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,6144,0.010462221999963125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,16384,0.020198222663667467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,4096,0.03980888923009237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,10240,0.022493332624435425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,32,0.012736000120639801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,32,0.0028133332315418455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,16384,0.09178666936026679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,7168,0.010965333216720156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,32,0.0029146667155954572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,768,0.0039013334446483185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,2560,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,65536,0.08158489068349202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,64,0.0028151110228565005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,3584,0.3367608918084039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,2048,0.012340444657537671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,12288,0.02882844540807936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,512,0.022798221972253587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,6144,0.12055467234717475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,128,0.014466666513019137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,6144,0.02973955538537767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,2560,0.020818667279349435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,8192,0.1376426617304484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,2048,0.05533688929345873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,768,0.01074044406414032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,3072,0.00889244427283605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,1024,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,65536,2.195486280653212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,16384,0.06002755297554863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,1536,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,768,0.007322666545708974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,2048,0.007978666987684038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,12288,0.1790168947643704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,7168,0.03849688834614224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,16384,0.033562666840023465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,3584,0.0074924445814556545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,3584,0.016168889072206285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,4096,0.020225778222084045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,16384,0.0388302206993103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,7168,0.008543999658690559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,10240,0.03314222229851617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,3584,0.007603555917739868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,1024,0.02236355510022905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,1024,0.004855111241340637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,64,0.006561777657932705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,5120,0.01758044461409251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,5120,0.030637333790461223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,1536,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,3072,0.007634667058785756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,256,0.015182221929232279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,256,0.012432888978057437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,5120,0.010062221851613786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,2048,0.01201244443655014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,128,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,256,0.008296888735559251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,7168,0.010856000085671743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,4096,0.20192178090413412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,32,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,128,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,3584,0.010709332923094431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,65536,0.15864622592926025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,16384,0.05077422327465481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,2048,0.023405333360036213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,7168,0.011547555526097616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,128,0.003198222153716617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,64,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,512,0.013792888985739814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,16384,0.03502577874395583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,128,0.015169777803950839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,1536,0.012019555601808759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,12288,0.1419502231809828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,6144,0.01670488880740272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,256,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,10240,0.2262479994032118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,6144,0.009737778041097853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,4096,0.007986666427718269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,3584,0.015526221858130561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,16384,0.046929776668548584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,1536,0.012420444852775998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,2560,0.013869333598348828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,16384,0.019520888725916546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,256,0.006863111009200414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,256,0.006702222343948152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,512,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,128,0.10481155580944485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,512,0.013772444592581855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,16384,0.03864533371395535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,8192,0.07434222433302137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,1024,0.004228444563017951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,5120,0.01276622215906779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,512,0.0032835555159383347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,8192,0.01643288963370853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,768,0.010507555471526252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,4096,0.017289777596791584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,10240,0.021336000826623704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,3072,0.016552888684802584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,3584,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,64,0.05443200137880114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,8192,0.04020266731580099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,6144,0.017895110779338412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,7168,0.016184000505341422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,10240,0.020648888415760465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,2048,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,7168,0.20492622587415907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,32,0.012806221842765808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,6144,0.10248800118764241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,512,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,5120,0.05676177475187513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,32,0.008802666432327693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,8192,0.37130223380194766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,16384,0.2176826662487454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,8192,0.04608711269166735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,2560,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,3584,0.01829600003030565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,3072,0.04850844542185465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,65536,0.038909332619773015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,2560,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,1024,0.006061333335108227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,6144,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,2560,0.007324444750944774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,768,0.016358221570650738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,2560,0.01553600033124288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,3584,0.006295111030340195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,1024,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,64,0.002833777831660377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,6144,0.21509244706895617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,65536,0.16305955251057944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,65536,0.0634746683968438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,768,0.015464888678656684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,2560,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,12288,0.014338667194048563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,256,0.017092444830470614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,512,0.013754666679435305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,10240,0.10993955532709758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,1024,0.02253155575858222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,2048,0.06728088855743408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,1024,0.042345778809653394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,32,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,64,0.002846222163902389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,8192,0.11033066776063706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,1536,0.004588444613748127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,512,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,12288,0.020776889390415616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,256,0.003170666595300039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,3584,0.011119999819331698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,3584,0.019374221563339233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,5120,0.12069955137040879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,4096,0.007771555748250749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,768,0.003556444413132138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,128,0.003859555555714501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,5120,0.007248000138335758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,512,0.0069164443347189165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,8192,0.010395555860466426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,16384,0.012472000386979846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,1024,0.004219555606444676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,3072,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,5120,0.023339556323157415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,5120,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,64,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,1024,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,65536,0.10395466619067723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,5120,0.06505155563354492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,3584,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,12288,0.016169778174824186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,3584,0.021431111627154883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,1024,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,128,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,6144,0.018917333748605516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,6144,0.03596622082922194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,64,0.002805333377586471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,6144,0.0099973330895106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,2560,0.05443111062049866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,4096,0.013733333183659447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,2048,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,1536,0.006063111126422882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,3584,0.007987555530336168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,65536,0.06379999717076619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,16384,0.026421333352724712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,65536,0.1459626621670193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,32,0.008622222476535374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,512,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,12288,0.0440995560752021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,4096,0.007607111500369177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,768,0.02659999993112352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,256,0.06844000021616618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,6144,0.009679111341635386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,768,0.012548444171746572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,7168,0.014831999937693277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,16384,0.042083554797702365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,1024,0.004295111116435793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,8192,0.01240799989965227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,4096,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,64,0.04209244582388136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,768,0.012886222038004132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,3072,0.01497244503762987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,2560,0.08218133449554443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,1536,0.11117066277398004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,7168,0.02477955487039354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,1024,0.00810933361450831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,2560,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,768,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,256,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,512,0.01569688816865285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,65536,0.1588968833287557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,6144,0.00828888929552502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,2048,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,2048,0.016827555166350473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,10240,0.01424533294306861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,1024,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,768,0.005117333183685939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,2048,0.08407733175489639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,4096,0.010985777609878115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,16384,0.02050933407412635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,3584,0.007662222617202335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,128,0.015689777003394235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,8192,0.02644888891114129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,3072,0.0125928885406918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,12288,0.04193688763512505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,7168,0.03407377666897244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,65536,0.06326844294865926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,64,0.2950364483727349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,2048,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,3584,0.010678222609890832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,7168,0.16842399703131783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,1536,0.01718488832314809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,128,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,2560,0.014519110321998596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,256,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,256,0.009407999614874521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,6144,0.028911110427644517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,3584,0.01071911139620675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,4096,0.04852888981501261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,128,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,10240,0.01551466683546702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,6144,0.010189333723651038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,65536,0.15833955340915257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,10240,0.013785777820481194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,256,0.01683911184469859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,768,0.017827555537223816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,2560,0.05934844414393107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,1536,0.0214373336897956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,5120,0.009369778136412302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,512,0.01351111133893331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,256,0.0123848890264829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,1536,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,3584,0.00793511089351442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,16384,0.05874933136834038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,512,0.0052382221652401825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,512,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,8192,0.012644444074895648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,2560,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,5120,0.00795022232664956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,5120,0.04790844519933065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,512,0.013152889079517789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,1024,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,128,0.011042666931947073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,12288,0.013433777623706393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,1024,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,3584,0.007928888830873702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,1024,0.041393776734670006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,12288,0.030544890297783747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,12288,0.031769778993394636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,12288,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,32,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,64,0.003583111282851961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,3072,0.006865777903132969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,8192,0.011049778097205691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,4096,0.02125866711139679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,1024,0.004237333519591225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,6144,0.00962400022480223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,2560,0.007326222128338284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,512,0.013069333301650153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,12288,0.025397333833906386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,1536,0.01407199932469262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,2560,0.12114222844441731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,1536,0.006244444598754247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,1536,0.015049777097172208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,65536,7.95633527967665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,5120,0.015139554937680563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,3584,0.08406044377221002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,10240,0.10689422157075669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,128,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,8192,0.017647110753589206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,65536,0.14028533299763998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,8192,0.02236000034544203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,1536,0.0281644430425432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,3584,0.018222222725550335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,6144,0.011699555648697747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,1536,0.013506666653686099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,7168,0.01074399964676963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,65536,0.10664621988932292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,128,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,5120,0.0069520001610120135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,65536,0.053605334626303784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,1536,0.020282665888468426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,3072,0.014832889040311178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,3072,0.0058648888435628675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,65536,0.08010133107503255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,5120,0.010454222559928894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,5120,0.011086222198274402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,16384,0.0833395587073432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,1536,0.005542222410440445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,3072,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,2560,0.010493333141009012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,3072,0.008631111019187504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,8192,0.023022222850057814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,256,0.015140444040298462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,512,0.018206222189797294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,1024,0.013371555341614617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,4096,0.04190222091144986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,65536,0.20885245005289713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,10240,0.0477751096089681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,16384,0.06462755468156603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,1536,0.024872889121373493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,5120,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,32,0.0042151109211974675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,2048,0.023730667101012334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,7168,0.124981337123447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,512,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,2560,0.02088800072669983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,512,0.006208000083764394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,4096,0.007652444144090016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,1536,0.06732800271775988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,32,0.008025777836640676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,8192,0.03500533435079787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,1536,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,512,0.005311111195219888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,256,0.003530666646030214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,8192,0.02075377768940396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,64,0.0031644445326593188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,4096,0.01686844395266639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,64,0.0047404443224271136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,256,0.005665777872006099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,4096,0.01990133358372582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,512,0.005741333381997214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,3584,0.007670222057236566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,2048,0.012605333493815528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,1536,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,65536,0.14001954926384821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,2048,0.01960088809331258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,6144,0.05067111055056254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,256,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,16384,0.04379466507169935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,3072,0.007924444145626491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,10240,2.9966390397813587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,16384,0.44687557220458984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,32,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,65536,0.06434755855136447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,128,0.002903999967707528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,2048,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,512,0.003912000192536248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,768,0.00387999994887246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,64,0.005669333454635408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,64,0.0033075554917256036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,16384,0.3975573380788167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,1536,0.007705777883529663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,3072,0.012539555629094442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,12288,0.15696088473002115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,1024,0.006016888966162999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,16384,0.05906133519278633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,1536,0.006992888947327931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,8192,0.010859555668301053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,64,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,6144,0.01998222205373976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,2560,0.017254221770498488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,8192,0.01891377733813392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,4096,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,1536,0.004543999830881755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,2048,0.009318222602208456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,64,0.0032106666929192017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,2048,0.016864000095261466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,512,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,2048,0.006334222025341458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,128,0.003156444471743372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,3584,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,64,0.003126222226354811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,3584,0.019215111931165058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,1024,0.005594666633341048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,64,0.009349333743254343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,8192,0.030352890491485596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,16384,0.008289777570300633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,65536,0.05645689037111071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,7168,0.010374222364690568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,10240,0.05277599891026815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,10240,0.10495199759801228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,256,0.017363554901546903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,128,0.0041911109454101985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,1024,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,4096,0.008671999805503422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,8192,0.012506666282812754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,3584,0.01856355534659492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,1536,0.016176889340082806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,6144,0.006575999988449945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,768,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,128,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,1536,0.013799111048380533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,2048,0.016502221425374348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,128,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,12288,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,3072,0.013202666408485837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,3072,0.023698666029506262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,256,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,16384,0.26103199852837455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,3072,0.011895111037625207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,7168,0.012039999994966718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,16384,0.03594933284653558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,2048,0.006908444480763541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,256,0.0038604442444112566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,5120,0.016680000556839835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,4096,0.024712888730896845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,2048,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,10240,0.015858666764365304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,3072,0.08615022235446507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,6144,0.011056000159846412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,7168,0.021144888467258875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,128,0.0052720002002186244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,8192,0.020757334099875558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,6144,0.009363555245929295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,4096,0.007311111523045435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,6144,0.022650667362742957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,7168,0.02036533256371816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,3072,0.03959644503063626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,8192,0.19365421930948892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,10240,0.014856888188256158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,512,0.02234933276971181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,8192,0.04318933354483711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,64,0.0032248888164758682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,1024,0.00527555536892679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,10240,0.05326844586266411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,1536,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,3072,0.007313778003056844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,64,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,6144,0.020244444410006206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,5120,0.013003555436929068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,12288,0.07183111376232572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,1024,0.024051555328898962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,5120,0.058851553334130176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,8192,0.04531911015510559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,6144,0.02387911081314087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,65536,0.1871617767545912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,16384,0.08449866374333699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,3072,0.02748888896571265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,64,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,512,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,2560,0.007670222057236566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,4096,0.022671111755900916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,768,0.015835555063353646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,8192,0.0299582216474745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,64,0.0029111111329661477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,512,0.012434666355450949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,8192,0.5971377690633138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,7168,0.02348800003528595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,4096,0.010220444036854638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,128,0.0026346666531430352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,65536,0.30808888541327584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,256,0.011057777537239922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,512,0.01479111115137736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,16384,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,2048,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,8192,0.15990044011010066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,2048,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,10240,0.008979555633332994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,3584,0.007623111208279927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,512,0.015569777952300178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,5120,0.008643555144468943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,12288,0.19647910859849718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,16384,0.049489776293436684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,16384,13.377587212456596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,3584,0.007313778003056844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,12288,0.03560622202025519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,7168,0.03087644444571601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,768,0.21845600340101454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,16384,0.34442933400472003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,4096,0.016908443636364408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,8192,0.010520888699425591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,6144,0.028488000233968098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,3584,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,5120,0.010029333333174387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,8192,0.0222053329149882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,512,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,768,0.009309333231714036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,12288,0.015853333804342482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,10240,0.019002666076024372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,128,0.012849777936935425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,1024,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,8192,0.010381333529949188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,3584,0.09509066740671794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,10240,0.025219556358125474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,512,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,1024,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,12288,0.018162666095627677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,3072,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,512,0.011533333195580376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,8192,0.07467199696434869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,5120,0.009453333086437648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,10240,0.02610133257177141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,16384,0.01976799964904785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,768,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,512,0.014129777749379476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,2560,0.014843554960356818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,12288,0.018204443984561496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,64,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,64,0.005574222240183089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,1024,0.018550222118695576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,8192,0.018719111879666645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,65536,0.20268800523546007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,6144,0.015496888094478183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,12288,0.01309333327743742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,768,0.018579555882347953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,1024,0.025212445192866858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,64,0.014072888427310519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,16384,0.05755644374423557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,3072,0.012874667015340594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,256,0.014379555980364481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,10240,0.01679377754529317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,1024,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,12288,0.04590577880541483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,1536,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,10240,0.016892444756295946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,768,0.016560888952679105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,3072,0.08989689085218643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,2048,0.00980533328321245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,64,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,1536,0.007010666860474481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,768,0.010065777434243096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,12288,0.6462800237867568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,65536,0.09593955675760905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,2048,0.016844444804721408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,7168,0.021852445271280076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,3072,0.1913315587573581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,512,0.042916443612840444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,512,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,512,0.013046222428480784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,768,0.01650399963061015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,3072,0.01386666629049513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,2560,0.024870221813519795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,64,0.008967999782827165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,1024,0.009476444787449306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,6144,0.012496888637542725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,1536,0.013056000073750814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,65536,0.06270222531424628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,5120,0.01663555536005232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,3584,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,1024,0.013074666261672974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,3584,0.015635554989178974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,5120,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,256,0.014436443646748861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,768,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,256,0.01643466618325975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,1536,0.009935110807418823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,8192,0.010858666565683154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,16384,0.053708444039026894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,3072,0.011754666765530905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,12288,0.03519377774662442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,768,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,32,0.00318755561278926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,7168,0.01959822244114346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,3584,0.058987557888031006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,8192,0.12096888489193386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,2560,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,128,0.015424888994958667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,64,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,16384,0.021366222037209406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,2560,0.007683555285135905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,1024,0.02456444501876831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,3584,0.011727111207114326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,256,0.027112889620992873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,7168,0.0601137810283237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,6144,0.019850666324297588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,3072,0.050703111622068614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,1536,0.007807999849319458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,5120,0.016992888516849942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,2560,0.13172977500491673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,4096,0.14631822374131945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,8192,0.016604445046848722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,7168,0.019254222512245178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,1024,0.014452444182501899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,2560,0.03383288780848185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,512,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,10240,0.012048888537618848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,512,0.01961866683430142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,12288,0.1703182193968031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,10240,0.015116444892353482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,64,0.0028613333900769553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,7168,0.010621333287821876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,3584,0.02334755493534936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,2048,0.018952000472280715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,512,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,2560,0.013195555243227216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,10240,0.022693332698610093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,65536,0.0506551100148095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,12288,0.1249671114815606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,6144,0.009047110875447592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,12288,0.025723555021815833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,4096,0.024283554818895128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,6144,0.008671999805503422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,128,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,10240,0.011432888607184092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,1024,0.007080888582600488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,2048,0.020344000723626878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,3584,0.010065777434243096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,3072,0.01622222198380364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,3072,0.015096000499195524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,7168,0.013737777868906656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,512,0.05901600254906548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,3072,0.03705599904060364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,256,0.016844444804721408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,128,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,512,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,65536,0.07795111338297527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,65536,0.15374933348761663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,768,0.012439111040698158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,2560,0.007999999655617608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,32,0.003228444399105178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,5120,0.01686755485004849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,16384,0.009715555442704095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,3584,0.005988444305128521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,256,0.003216000066863166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,7168,0.02035199933581882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,1024,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,1024,0.0075911109646161394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,4096,0.0304115547074212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,5120,0.01688088807794783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,3584,0.009543110926946005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,1024,0.27514489491780597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,6144,0.131331549750434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,5120,0.01938577824168735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,256,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,3584,0.023316444622145757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,8192,0.16265866491529676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,2560,0.0162124451663759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,512,0.0032231110251612137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,16384,0.02679644525051117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,1536,0.02840977907180786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,10240,0.011393778026103973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,3584,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,7168,0.02095199955834283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,10240,0.043262223402659096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,2048,0.009081777599122789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,2048,0.009012444151772393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,768,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,256,0.014841778410805596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,5120,0.008644444247086843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,8192,0.020977778567208186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,2048,0.03792444533771939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,3072,0.045180443260404796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,16384,0.02401955591307746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,3072,0.010183999935785929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,3072,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,5120,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,32,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,4096,0.025054223007626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,16384,0.0070062221752272705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,32,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,16384,0.015808888607554965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,3584,0.029122667180167303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,2048,0.037393778562545776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,2560,0.01595288846227858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,12288,0.014087999860445658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,12288,0.01349688900841607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,2560,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,768,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,64,0.0028533333291610083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,8192,0.025204444924990337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,10240,0.012720888687504662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,7168,0.12558221817016602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,1536,0.01311911145846049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,256,0.016813332835833233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,65536,2.069825702243381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,4096,0.011721777419249216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,128,0.022295110755496558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,3072,0.0425146652592553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,3072,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,4096,0.041301333241992526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,7168,0.010683555569913654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,4096,0.0079644446571668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,6144,0.01276711126168569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,512,0.0042053332759274375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,1024,0.008802666432327693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,1024,0.004531555705600315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,65536,0.14389867252773708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,4096,0.024825778272416856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,256,0.006654222392373615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,256,0.005255110975768831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,7168,0.009675555759006077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,256,0.0038515557017591265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,1536,0.008949333594905006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,256,0.0031831111345026228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,2560,0.031147556172476873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,4096,0.1339368952645196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,7168,0.023033777872721355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,2560,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,16384,0.014175110393100314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,10240,0.011388444238238864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,7168,0.01962577799956004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,5120,0.04163644380039639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,64,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,64,0.003875555677546395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,1536,0.013804444836245643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,2560,0.007281777759393056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,1024,0.0042035554846127825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,2560,0.00758844448460473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,1024,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,5120,0.0138568886452251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,768,0.01239466667175293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,4096,0.013800000150998434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,1024,0.005772444523043103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,65536,0.1035191085603502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,10240,0.01686755485004849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,6144,0.6993075476752387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,4096,0.08549244536293878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,7168,0.02498400045765771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,16384,0.029102222787009344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,1536,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,768,0.01716888944307963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,8192,0.010392888552612728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,64,0.003452444449067116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,256,0.022335999541812476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,1536,0.00620888877246115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,5120,0.008664888640244802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,5120,0.027056889401541814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,12288,0.02367111047108968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,2560,0.0787022246254815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,1024,0.014456000592973499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,16384,0.016365332735909354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,6144,0.012166221936543783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,6144,0.9013901816474067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,16384,0.424544890721639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,768,0.007734222544564142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,2560,0.010464000205198923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,10240,0.026625777284304302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,8192,0.0148062225845125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,16384,0.03570311268170675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,7168,0.035495112339655556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,65536,0.13766488764021131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,1536,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,10240,0.029711110724343195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,64,0.0029057777590221832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,128,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,1536,0.006350222147173352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,64,0.002806222273243798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,2048,0.01757600075668759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,16384,0.019021333919631112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,256,0.006243555496136348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,768,0.003564444267087512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,768,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,16384,0.015106666419241162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,2560,0.006681777950790193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,5120,0.025452444950739544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,3584,0.07068444622887506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,32,0.005991111199061076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,768,0.004724444614516364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,5120,0.02715644405947791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,10240,0.07684177822536893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,2048,0.0063324446479479475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,12288,0.03436977664629618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,65536,0.11995822853512234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,16384,0.3277448813120524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,128,0.04364888866742452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,256,0.03992799917856852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,768,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,768,0.005800888770156437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,1536,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,512,0.003165333428316646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,2560,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,6144,0.0844773319032457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,5120,0.15176267094082302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,8192,0.01071466671095954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,7168,0.009697777529557547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,12288,0.03671377897262573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,8192,0.010695999695195092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,12288,0.38217777676052517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,65536,11.540939331054688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,1024,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,1024,0.018242667118708294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,768,0.023015111684799194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,512,0.003216888962520493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,512,0.012049777640236748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,2560,0.06419289112091064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,8192,0.14630844857957628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,7168,0.029088000456492107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,2048,0.008357333640257517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,16384,0.042203555504480995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,256,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,8192,0.2230186727311876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,5120,0.059598220719231494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,16384,0.03678755627738105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,12288,0.06393688917160034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,6144,0.011695110963450538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,8192,0.1156275537278917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,128,0.0038444445365005066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,2560,0.009045333498054082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,2048,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,10240,0.11019111341900295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,8192,0.0140328887436125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,10240,0.01165244479974111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,128,0.010384889112578498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,6144,0.03419555558098687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,2048,0.007895111209816402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,2560,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,128,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,256,0.003622222277853224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,16384,0.012834666503800286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,10240,0.015288889408111572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,1536,0.04063911239306132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,4096,0.011131555669837527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,8192,0.024267555938826665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,1536,0.006048000107208888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,65536,0.22102133433024088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,1536,0.1292382213804457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,65536,0.012707555459605323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,3584,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,256,0.006139555739031897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,8192,0.016203555795881484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,768,0.003560888684458203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,64,0.00565422202150027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,2560,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,7168,0.009979555176364051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,512,0.028091556496090356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,10240,0.08204178015391032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,12288,0.013064000341627332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,16384,0.379423114988539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,10240,0.02474844455718994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,768,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,128,0.0053022222386466126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,4096,0.00795199970404307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,64,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,2048,0.017112000121010672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,512,0.00830488900343577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,64,0.008999110923873054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,7168,0.010412444670995077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,256,0.005677333308590784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,5120,0.0124177775449223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,64,0.007319110963079665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,2048,0.006608888920810487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,128,0.007655111451943715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,3584,0.008214222060309516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,32,0.007176000210973952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,8192,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,256,0.009496000077989366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,768,0.004090666770935059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,512,0.003334222154484855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,128,0.0053137776752312975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,1536,0.006340444501903322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,5120,0.02619022296534644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,32,0.0028133332315418455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,10240,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,2048,0.01942133406798045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,4096,0.015988444288571674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,10240,0.026825777358478967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,2560,0.014108444253603617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,2048,0.05858844518661499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,3584,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,16384,0.025574222207069397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,4096,0.00794488853878445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,1024,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,2048,0.008769777913888296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,6144,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,16384,0.03577333357599046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,16384,0.03135200010405646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,1536,0.060285336441463895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,2048,0.018214222457673814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,256,0.015180443723996481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,10240,0.011878222227096558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,768,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,2560,0.01684711045689053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,768,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,768,0.013047110703256396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,10240,0.01035199976629681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,4096,0.00796088907453749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,3584,0.08934577968385483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,5120,0.008432000047630733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,512,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,64,0.005983110931184556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,3584,0.041146665811538696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,12288,0.01130488928821352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,65536,0.3006062242719862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,2048,0.013210666676362356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,12288,0.04558666547139486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,2048,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,5120,0.01995111174053616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,3584,0.008852444589138031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,256,0.005629333357016246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,5120,0.010774222513039907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,256,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,64,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,1536,0.01457333399189843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,12288,0.02427644365363651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,1536,0.010397333237859938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,768,0.10813066694471572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,1024,0.0041662222809261745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,4096,0.01756622228357527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,12288,0.03195110956827799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,4096,0.015113777584499784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,65536,0.10061511066224839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,3072,0.008027555214034187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,3072,0.05337511168585884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,65536,0.08966844611697727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,768,0.006205333189831839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,16384,0.019875556230545044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,1024,0.013804444836245643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,256,0.015097777048746744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,512,0.005993777679072485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,6144,0.025809778107537165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,64,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,256,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,1024,0.017941332525677152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,12288,0.10648266474405925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,3584,0.00906666699382994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,2048,0.027127999398443434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,16384,0.16442488299475774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,256,0.0031617778456873367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,65536,0.3214915593465169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,7168,0.05253333184454176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,12288,0.028485334581798975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,512,0.004561777744028303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,768,0.005548444473081165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,3584,0.007970666719807519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,256,0.011214222345087262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,2560,0.007280000381999546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,1536,0.01622488929165734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,65536,0.11286221610175239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,10240,0.01632000009218852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,6144,0.28907643424140084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,512,0.012855999999576144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,768,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,512,0.006000888844331105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,2048,0.006691555596060223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,6144,0.018195556269751657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,512,0.009582222335868413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,7168,0.010140444669458602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,6144,0.012542222109105853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,16384,0.040292445156309344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,3072,0.01995111174053616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,1536,0.023792889383104112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,2560,0.0073520003093613525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,65536,0.014280888769361707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,512,0.013419555293189155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,65536,0.303056001663208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,8192,0.03225333491961161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,6144,0.024288889434602525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,5120,0.029507554239696924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,5120,0.19635644223954943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,64,0.020408888657887776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,768,0.005384888913896348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,2048,0.012375999655988483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,5120,0.011020444333553314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,2048,0.006877333339717653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,4096,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,256,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,16384,0.023976888921525743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,5120,0.04846133457289802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,16384,0.0673475530412462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,256,0.015467555986510383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,10240,0.026905778381559584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,768,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,1024,0.00870577742656072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,8192,0.049327999353408813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,65536,0.04900177650981479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,1024,0.0038524443904558816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,4096,0.02090222140153249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,256,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,10240,0.02808177802297804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,4096,0.020512888828913372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,3072,0.09250489208433364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,3072,0.2245244450039334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,65536,0.08099200328191121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,10240,0.015882665912310284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,8192,0.021746666895018682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,2560,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,2048,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,6144,0.019647111495335896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,12288,0.022337777747048274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,10240,0.029571556382709082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,768,0.01275199982855055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,65536,0.05626755290561252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,64,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,768,0.024350222614076402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,2560,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,12288,0.01684533390733931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,1536,0.004558222161398994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,768,0.01684799955950843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,7168,0.013596444494194455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,5120,0.00866844422287411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,3584,0.009045333498054082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,65536,0.07515377468532987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,5120,0.008683555656009251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,7168,0.04225066635343763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,1536,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,12288,0.029126223590638902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,768,0.003671110918124517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,65536,0.16422222720252141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,1536,0.007982222570313348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,512,0.02478844424088796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,65536,0.1503022246890598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,10240,0.02234577801492479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,2560,0.029940442906485662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,65536,0.143103109465705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,768,0.011722666521867117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,64,0.003136000078585413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,10240,0.1553315586513943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,512,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,5120,0.047480887836880155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,3072,0.018400000201331247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,6144,0.015807110402319167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,3584,0.017222222354676988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,4096,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,512,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,2560,0.007520889242490132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,2560,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,65536,0.07850755585564508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,256,0.006879111131032308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,1536,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,256,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,16384,0.025360888904995386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,256,0.015821332732836407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,1024,0.029742221037546795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,4096,0.008745777938101027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,128,0.005261333452330695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,65536,0.08361066712273492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,1536,0.023405333360036213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,4096,0.016933333542611863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,5120,0.011057777537239922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,64,0.0028702221396896574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,2560,0.016884444488419425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,65536,0.04525688952869839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,6144,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,6144,0.022988445229000513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,7168,0.2323991192711724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,3584,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,3584,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,512,0.014459555347760519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,7168,0.01367911116944419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,12288,0.0063688887490166565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,6144,0.05614933040406969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,3584,0.01444533301724328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,4096,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,6144,0.009005332986513773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,7168,0.029437333345413208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,768,0.017864889568752713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,512,0.00351200004418691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,8192,0.02992977698644002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,7168,0.030280888080596924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,10240,0.012252444194422828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,7168,0.010073777702119615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,7168,0.01036177741156684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,8192,0.013075555364290873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,6144,0.013099555340078143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,3584,0.007657777931955125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,4096,0.011373333632946014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,512,0.012791111237472959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,7168,0.0665502217080858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,2048,0.042288889487584434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,10240,0.01719555589887831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,65536,0.3894302315182156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,256,0.00526311124364535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,4096,0.09560266468260024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,16384,0.1593759987089369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,16384,0.03875111209021674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,64,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,256,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,128,0.0028364445186323593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,64,0.024910221497217815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,65536,0.06185955471462674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,768,0.0053280000057485365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,768,0.003751111113362842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,10240,0.035381333695517644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,128,0.015086222026083203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,12288,0.042091555065578885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,6144,0.01908177799648709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,2048,0.014377777775128683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,8192,0.015504888362354703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,3072,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,6144,0.019588443968031142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,256,0.00490133340160052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,128,0.00656088896923595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,256,0.009317333499590555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,128,0.0030053332448005676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,1536,0.007623111208279927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,64,0.005554666535721884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,7168,0.012853333519564735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,8192,0.01540177729394701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,65536,0.06510666343900892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,10240,0.14288532733917236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,7168,0.010048000348938836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,12288,0.018100443813535903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,768,0.015511110424995422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,8192,0.02719288898838891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,6144,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,10240,0.04965510964393616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,4096,0.009355555805895064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,128,0.023003555006451074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,12288,0.03615466753641764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,3584,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,8192,0.03887644410133362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,3584,0.0562737782796224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,768,0.004190222256713443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,2048,0.12089599503411187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,3584,0.020982222424613103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,64,0.012392889294359418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,65536,6.640343136257595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,3072,0.020595555504163105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,2560,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,1536,0.018575999471876357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,4096,0.008012444608741337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,2048,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,1536,0.013812444276279874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,16384,0.04645244280497233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,8192,0.01956977778010898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,128,0.016226665841208566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,3072,0.01756355497572157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,32,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,2560,0.005670222143332164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,2048,0.025012445118692186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,8192,0.0110871113008923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,6144,0.017768889665603638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,16384,0.22204177909427217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,12288,0.07355555560853747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,1024,0.018921777606010437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,16384,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,768,0.014454222387737699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,768,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,1024,0.01571377780702379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,768,0.017515555024147034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,2560,0.027335110637876723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,10240,0.014750222365061441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,64,0.00673333348499404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,768,0.015484443969196744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,512,0.013417777915795645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,65536,0.1486577722761366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,2048,0.07192266649670072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,16384,0.01591022147072686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,65536,0.061250666777292885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,3072,0.07997955878575642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,2048,0.008737777670224508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,2048,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,1536,0.0052151112920708126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,10240,0.1907271146774292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,512,0.003228444399105178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,4096,0.09097599983215332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,5120,0.01440533333354526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,768,0.004363555461168289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,3584,0.014099554883109199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,128,0.015170666906568738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,2560,0.009863999982674917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,1536,0.0074782222509384155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,32,0.002728888971938027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,8192,0.012969777815871768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,7168,0.009988444546858469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,128,0.0028400001012616684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,12288,0.04313866628540886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,65536,0.09940177864498562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,4096,0.013878222141000958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,65536,1.2741858164469402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,2048,0.016697777642144095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,3072,0.030124445756276447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,3072,0.007335111498832703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,3072,0.015524443652894763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,4096,0.05006666647063362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,2048,0.014397333065668741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,512,0.06143110990524292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,128,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,4096,0.12511644098493788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,8192,0.014270222849316068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,1024,0.009025777379671732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,32,0.019985778464211356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,5120,0.023815999428431194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,4096,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,6144,0.013783999615245394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,32,0.009351111120647853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,2560,0.01904533306757609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,2560,0.019567999574873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,768,0.0188071115149392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,65536,0.07814577553007337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,2560,0.009342222577995723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,65536,0.9153529273139106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,2048,0.01346666696998808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,8192,0.01849688920709822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,7168,0.010034666293197209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,768,0.003528888854715559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,3584,0.010993777877754636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,2560,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,10240,0.03139644530084398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,12288,0.09567555454042222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,12288,0.1046524445215861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,10240,0.007924444145626491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,5120,0.06088978052139282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,6144,0.008338666624493068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,3584,0.08848000235027736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,1536,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,6144,0.1210062238905165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,8192,0.02036177780893114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,128,0.007967111137178209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,5120,0.011154666543006897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,64,0.0031662223239739737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,10240,0.01960000064637926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,16384,0.036335110664367676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,256,0.014868444866604276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,65536,0.062464886241488986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,12288,0.018975999620225694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,128,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,3584,0.018578666779730055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,16384,0.042486220598220825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,16384,0.05403288867738512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,8192,0.025435555312368605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,10240,0.016612443659040663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,64,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,3584,0.02057422200838725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,64,0.0032471112079090546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,768,0.004536889079544279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,32,0.0028924445311228433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,5120,0.014121777481502958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,1536,0.006232000059551663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,256,0.0031111112071408164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,12288,0.016855110724767048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,65536,0.1765813297695584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,128,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,3072,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,768,0.004253333227501975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,8192,0.02161866592036353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,4096,0.00794666674402025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,10240,0.08509866396586101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,2560,0.013632888595263163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,7168,0.1598337756262885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,512,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,2048,0.013781333135233985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,3584,0.013580444786283704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,5120,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,6144,0.00828888929552502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,256,0.0042053332759274375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,768,0.0058648888435628675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,16384,0.017887999614079792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,2560,0.015459555718633862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,1536,0.024875554773542616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,65536,0.07035377952787611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,6144,0.013857777747843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,16384,0.10872711075676812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,8192,0.0105262224872907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,5120,0.00906133320596483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,128,0.007825777762466008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,16384,0.020051555501090158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,7168,0.03263555632697211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,2048,0.019582221905390423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,768,0.003823111040724648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,8192,0.16798044575585258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,65536,0.08312888940175374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,8192,0.028794666131337483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,3072,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,4096,0.010899555351999072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,4096,0.06641422377692328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,10240,0.030231999026404485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,128,0.008731555607583787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,32,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,8192,0.0707191096411811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,4096,0.02053600052992503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,2560,0.1900177796681722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,4096,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,65536,0.0680275559425354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,4096,0.016252444850073922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,7168,0.01973600023322635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,1024,0.023952888117896184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,12288,0.013357333011097379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,2560,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,2048,0.014510222607188754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,5120,0.03158933255407546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,10240,0.042912890513738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,12288,0.04273510972658793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,3072,0.06375822093751696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,2560,0.04704533351792229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,3072,0.0069857777820693115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,12288,0.01979022224744161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,10240,0.10756622420416938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,6144,0.02808977829085456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,10240,0.1537733342912462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,256,0.0052151112920708126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,3072,0.011038222246699862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,768,0.029389331738154095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,7168,0.030568887790044148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,8192,0.034373333056767784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,512,0.008374222450786168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,8192,0.01166488892502255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,32,0.0068400001360310455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,6144,0.010662222074137794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,1024,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,1536,0.015503111812803479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,3072,0.009570666485362584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,6144,0.038943999343448214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,6144,0.0099982221921285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,32,0.002817777709828483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,3072,0.03135733471976386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,12288,0.13514667087131077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,16384,0.30322310659620494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,8192,0.011184000306659274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,10240,0.2085955540339152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,12288,0.02573066618707445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,65536,0.07499288850360446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,1536,0.01308444473478529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,65536,0.056863109270731606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,768,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,65536,0.07865333557128906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,128,0.0029964444951878656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,2048,0.006249777972698212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,65536,0.11894844637976752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,2560,0.029764443635940552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,2560,0.08555289109547932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,3584,0.07769778039720324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,512,0.007346666521496243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,16384,0.014544000228246054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,16384,0.019725332657496136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,7168,0.023615999354256526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,3072,0.019283556275897555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,1024,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,128,0.003321777822242843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,16384,0.048842665221956044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,6144,0.02204977803760105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,128,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,768,0.004910222358173794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,6144,0.021944888763957556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,12288,0.10555999808841282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,8192,0.013917333549923368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,2560,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,1536,0.01275466630856196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,1536,0.004271111140648524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,5120,0.0925253364774916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,65536,9.010223388671875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,768,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,7168,0.009552888572216034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,7168,0.015863110621770222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,7168,0.01680444512102339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,3584,0.07045777638753255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,1536,0.006728000111050076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,32,0.006989333364698622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,6144,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,12288,0.008569777839713627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,2560,0.07782755295435588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,65536,0.06945866346359253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,2048,0.015899555550681222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,12288,0.013146667016877068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,4096,0.02472711106141408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,768,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,65536,0.06569333208931817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,5120,0.009352000223265754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,4096,0.008007110820876228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,10240,0.011575111084514193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,7168,0.016629333297411602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,256,0.004570666700601578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,10240,0.02572444412443373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,256,0.003552000141806073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,64,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,5120,0.009380444056457942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,128,0.029222223493787978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,3584,0.009002666506502364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,2560,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,3584,0.024845333562956914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,65536,0.20334577560424805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,128,0.0028853333658642242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,1024,0.010938666760921478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,8192,0.015807110402319167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,3584,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,3072,0.016564443707466125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,1024,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,3584,0.02199111051029629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,12288,0.02882755464977688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,768,0.016866667403115165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,3072,0.11870133876800537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,4096,0.007669332954618666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,256,0.003168888803985384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,65536,0.13276799519856772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,10240,0.06958044237560697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,128,0.0057839999596277876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,768,0.012807110945383707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,1024,0.016508445143699646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,16384,0.051479998562071055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,4096,0.007948444121413762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,2048,0.02407288882467482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,5120,0.43980177243550617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,6144,0.017647110753589206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,3584,0.01204711116022534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,16384,0.15759644243452284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,1536,0.007709333466158972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,64,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,7168,0.01000533335738712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,2560,0.009343110852771336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,32,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,6144,0.013646221823162503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,12288,0.02994933393266466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,12288,0.12866844071282282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,5120,0.014119111829333834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,512,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,3584,0.01787911189926995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,64,0.015861334072219003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,2560,0.061800890498691134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,16384,0.048156443569395274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,1536,0.004607111215591431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,64,0.0028817777832349143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,8192,0.010041777458455827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,3072,0.021254221598307293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,2560,0.007996444072988298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,2560,0.020983111527231004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,8192,0.02130488885773553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,2560,0.016532444291644625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,512,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,7168,0.009534222384293875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,65536,0.06672710842556424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,8192,0.030654221773147583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,256,0.0034933334423436057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,16384,0.026911111341582403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,2560,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,3584,0.007979555262459649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,6144,0.01735733283890618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,16384,0.060977776845296226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,2560,0.015783111254374187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,12288,0.05513155460357666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,768,0.01512977812025282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,10240,0.026737777723206416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,16384,0.043813334570990674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,128,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,512,0.014092443717850579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,1024,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,7168,0.013418667018413544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,4096,0.02368444369898902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,128,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,5120,0.008965333302815756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,12288,0.014540443817774454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,2048,0.020255999432669747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,128,0.014490667316648694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,2560,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,8192,0.051573332813051015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,8192,0.025026667449209426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,10240,0.04266488883230421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,64,0.008047999607192146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,65536,0.20645422405666777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,32,0.0028986665937635633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,10240,0.01944266590807173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,5120,0.009548444714811113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,4096,0.022856889499558344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,7168,0.02104088829623328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,12288,0.023702222439977858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,64,0.003526222374704149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,12288,0.028770665327707928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,3072,0.008712000317043727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,65536,0.08870666556888157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,128,0.026941332552168105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,128,0.01481066644191742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,256,0.0038480001191298165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,5120,0.014716444744004143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,2560,0.008623999853928884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,65536,0.08735999796125625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,32,0.005635555419656966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,7168,0.021871111459202234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,2048,0.01613333324591319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,7168,0.020640888147883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,6144,0.006090666684839461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,2048,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,2048,0.01422755585776435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,2048,0.061979558732774526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,2560,0.02239111065864563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,128,0.01521511044767168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,5120,0.03479289015134176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,8192,0.017980444762441847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,2560,0.008457777400811514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,2560,0.007280000381999546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,8192,0.01578044394652049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,512,0.010260444548394945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,768,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,512,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,32,0.005215999980767568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,5120,0.021426666114065383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,7168,0.012220444778601328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,128,0.044052445226245456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,256,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,8192,0.010398222340477837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,16384,0.04794755578041077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,3584,0.008632889224423302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,7168,0.015048000547620984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,64,0.008273777862389883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,64,0.0029057777590221832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,2048,0.04716089036729601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,2048,0.01513066722287072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,12288,0.013520888984203339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,7168,0.01830222209294637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,6144,0.009748444788985783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,64,0.0029155556112527847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,10240,0.012103999654452005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,2560,0.005803555664088991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,1024,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,2048,0.012221333053376941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,4096,0.008741333252853816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,2560,0.026103999879625108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,512,0.02481777800454034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,6144,0.014761777387724983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,768,0.005385777602593104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,12288,0.05146577623155382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,2048,0.006615110983451207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,768,0.016231111354298063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,2560,0.006872889068391588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,6144,0.012895999683274163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,4096,0.00758133331934611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,768,0.00665511108107037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,16384,0.03218844532966614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,10240,0.059657779004838735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,128,0.014502222339312235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,32,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,7168,0.013760888742076026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,4096,0.016799110505315993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,8192,0.04446666770511203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,2048,0.04555111130078634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,3584,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,5120,0.008929777476522658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,10240,0.03028977910677592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,2048,0.02846577763557434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,6144,0.20702132913801405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,10240,0.026131555438041687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,3072,0.019007111589113872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,128,0.0028479999552170434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,128,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,4096,0.020353777541054618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,5120,0.02032355633046892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,1024,0.009391999906963771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,5120,0.023797333240509033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,5120,0.01201866649919086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,16384,0.15403110451168484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,512,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,768,0.023258666197458904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,65536,0.20731377601623535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,4096,0.014856000741322836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,64,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,256,0.005655999812814925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,2048,0.006530666516886816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,6144,0.008007110820876228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,1536,0.013403555585278405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,5120,0.020603555772039626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,2048,0.006696888970004187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,6144,0.009360888765917884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,6144,0.07849511173036364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,2048,0.0070222218831380205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,5120,0.01828888886504703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,12288,0.6141529083251953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,1536,0.022312000393867493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,512,0.10755733648935954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,65536,0.08360355430179173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,65536,0.1859804391860962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,12288,0.040714667903052434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,65536,0.037159111764695905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,7168,0.012808000048001608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,64,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,64,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,8192,0.19814666112263998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,10240,0.0140364451540841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,12288,0.017426666286256578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,5120,0.012392889294359418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,4096,0.03758666581577725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,128,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,3072,0.01034044474363327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,5120,0.02161955502298143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,10240,0.030396444929970637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,12288,0.021179555190934077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,10240,0.10318489207161798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,32,0.020270221763186984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,3072,0.027007111244731482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,4096,0.052207112312316895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,3072,0.007663999994595845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,768,0.010713777608341642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,256,0.003253333270549774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,1024,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,128,0.0029022221763928733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,64,0.005873777800136142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,1536,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,1024,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,2560,0.0069644442862934535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,256,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,5120,0.03524355424775018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,512,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,1024,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,3584,0.030336889955732558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,7168,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,12288,0.12995289431677923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,10240,0.03605511122279697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,256,0.02035466664367252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,128,0.006074666563007567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,4096,0.22547290060255262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,8192,0.02606666584809621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,4096,0.026642666922675237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,768,0.016876444220542908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,7168,0.019083556201722886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,3072,0.09892621967527603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,6144,0.016601777738995023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,4096,0.008973333570692275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,12288,0.0296053323480818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,12288,0.029804444975323145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,12288,0.023542222049501207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,3072,0.010063110954231685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,16384,0.02091377807988061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,3072,0.10204266839557224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,768,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,5120,0.010122666756312052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,3584,0.018035555879275005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,6144,0.022358222140206233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,1536,0.009022222624884712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,3584,0.01661155621210734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,3072,0.013723555538389417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,2560,0.011582222249772815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,32,0.006301333506902059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,4096,0.18785066074795195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,8192,0.015015999476114908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,2560,0.006309333360857434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,128,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,6144,0.02459555533197191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,512,0.031680001152886286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,1536,0.014456000592973499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,2560,0.015455111861228943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,8192,0.020248888267411124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,64,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,10240,0.047889779011408486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,256,0.003206222214632564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,8192,0.011083555718262991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,2048,0.010400888820489248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,4096,0.018968888454967074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,65536,0.3206515577104357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,3584,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,12288,0.0623191131485833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,12288,0.024022221565246582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,512,0.005489777773618698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,1024,0.004190222256713443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,2048,0.016598222984208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,5120,0.00610133343272739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,512,0.007183999651008182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,64,0.005720888988839255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,512,0.004335111214054955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,256,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,256,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,512,0.01408088869518704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,3584,0.010657777388890585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,2048,0.008009778128729926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,768,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,7168,0.0558240016301473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,6144,0.0189057770702574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,6144,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,10240,0.011448889142937131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,1536,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,1536,0.03211822112401327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,16384,0.017097777790493436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,3072,0.011785777906576792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,6144,0.018927999668651156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,768,0.0052151112920708126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,10240,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,2048,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,7168,0.014164444473054675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,2048,0.008343111309740279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,3072,0.010767111347781287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,4096,0.009305777649084726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,16384,0.03965866565704346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,256,0.005382222019963794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,2560,0.011743110915025076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,3584,0.006231110956933763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,6144,0.0088755554623074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,4096,0.19410044617123076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,12288,0.0221368885702557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,256,0.0063182223174307085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,768,0.041548444165123835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,32,0.018914666440751817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,3584,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,512,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,8192,0.025156444973415796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,3072,0.020776000287797716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,16384,0.06764177481333415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,1024,0.017120000388887193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,1024,0.004543111142185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,16384,0.010069333016872406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,1536,0.03937777876853943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,128,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,2048,0.006956444846259222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,12288,0.180367112159729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,128,0.02756533357832167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,16384,0.10073422061072455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,10240,0.06191644403669569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,10240,0.012103111379676394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,128,0.004204444587230682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,10240,0.01180266671710544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,3584,0.008842666943868002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,7168,0.010779555473062726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,65536,0.10338844193352593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,10240,0.021779555413458083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,4096,0.011031111081441244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,128,0.0052266667286554975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,65536,0.04538666539722019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,1536,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,6144,0.12433955404493545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,8192,0.016454221473799813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,1536,0.008632889224423302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,65536,0.06398489077885945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,8192,0.1677751143773397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,7168,0.053207112683190234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,2048,0.020488889680968392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,8192,0.016605334149466622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,2560,0.0070186663005087115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,12288,0.01447111037042406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,256,0.02478933334350586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,1536,0.006550222221348021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,128,0.03314933180809021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,1536,0.009662222531106737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,7168,0.03437422381507026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,10240,0.00979200005531311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,65536,0.16605422231886122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,2560,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,16384,0.09026488992902969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,1024,0.03572533196873135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,4096,0.028388443920347426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,6144,0.05652533637152778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,4096,0.01956977778010898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,3072,0.031133333841959637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,1536,0.0063182223174307085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,3584,0.011647111011876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,4096,0.009093333449628618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,7168,0.012056888805495368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,2560,0.041328890456093684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,12288,0.01551644504070282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,3584,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,2560,0.09260711405012344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,1024,0.004879999905824661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,2560,0.010240000155236986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,128,0.01603555513752831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,12288,0.013445333474212222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,12288,0.03148622314135233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,768,0.006073777874310811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,2560,0.01053600013256073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,6144,0.0090844440791342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,12288,0.05059377683533562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,128,0.0033484444850020935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,12288,0.03872088922394647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,65536,0.14982933468288845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,10240,0.017687999539905123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,6144,0.009499555660618676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,5120,0.029924445682101782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,3072,0.006232888748248418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,1024,0.016932444439993966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,8192,0.008276444342401292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,1536,0.04257333278656006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,1024,0.012119111087587146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,1536,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,1536,0.015469332536061605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,7168,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,256,0.22003377808464897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,32,0.003519111209445529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,3584,0.008632000121805403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,128,0.014442667365074158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,1024,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,6144,0.009010666774378883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,3072,0.10204533073637222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,2048,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,4096,0.010494222243626913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,2048,0.012913777596420713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,3584,0.007353777686754863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,3584,0.00794577764140235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,5120,0.016620443926917184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,8192,0.028560890091790095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,1536,0.026924444569481745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,10240,0.011708444191349877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,64,0.006588444527652528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,7168,0.01108000013563368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,128,0.007269333634111617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,3072,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,65536,0.07377511262893677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,5120,0.008169777691364288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,768,0.003871110992299186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,6144,0.012030222349696688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,512,0.006990222467316522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,256,0.013076444466908773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,256,0.008295999632941352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,5120,0.01684799955950843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,32,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,10240,0.025256888733969793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,128,0.01847466660870446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,5120,0.024511110451486375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,7168,0.027609777119424608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,2560,0.029907554388046265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,10240,0.037159999211629234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,2048,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,3584,0.030805332793129817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,7168,0.014293332894643148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,1024,0.006362666686375936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,65536,0.07198044326570299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,1536,0.011867555479208628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,5120,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,12288,0.09612711270650227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,6144,0.027066666218969557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,2048,0.010079999764760336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,16384,0.21726489067077637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,1024,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,6144,0.013041777743233575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,6144,0.02570666703912947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,3584,0.010661332971519895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,2048,0.00979377743270662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,1536,0.01587733295228746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,10240,0.3238106568654378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,3584,0.042768889003329806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,256,0.058583113882276744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,12288,0.030262221892674763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,7168,0.02474400069978502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,768,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,2560,0.007271111011505127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,10240,0.07974400122960408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,3072,0.20211288664076063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,8192,0.02755733331044515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,3584,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,5120,0.013736888766288757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,6144,0.01220355596807268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,8192,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,2048,0.010999110837777456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,65536,0.1488355530632867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,512,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,8192,0.0277128881878323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,128,0.015397333436542086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,1024,0.02590933276547326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,1536,0.022608889473809138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,12288,0.013451555536852943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,512,0.02307022280163235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,512,0.015505777464972602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,128,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,3584,0.07804177867041694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,128,0.01585422290696038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,7168,0.04184177849027845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,5120,0.00869422240389718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,6144,0.06836977932188246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,7168,0.014653333359294467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,12288,0.046600000725852124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,2560,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,64,0.05325955483648512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,1024,0.006048888795905643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,5120,0.016677333248986136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,8192,0.011410666836632622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,4096,0.020020445187886555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,512,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,512,0.004936000125275718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,6144,0.009376889301670922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,2560,0.009365333451165093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,8192,0.009301332963837517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,5120,0.01275288893116845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,6144,0.038041777080959745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,128,0.002826666666401757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,2048,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,6144,0.012568000290128918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,768,0.03727466530270047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,64,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,3584,0.013550221920013428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,12288,0.05017333229382833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,8192,0.17852266629536948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,4096,0.008021333151393467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,10240,0.022450667288568284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,16384,0.021104888783560857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,5120,0.021474666065639917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,32,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,3072,0.007391999993059371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,32,0.0032826666202810076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,256,0.19729688432481554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,2048,0.030993776188956365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,5120,0.046021332343419395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,768,0.00999466660949919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,10240,0.027248889207839966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,65536,0.15308533774481878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,1536,0.021458667185571458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,6144,0.011718221836619906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,10240,0.07826844188902113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,5120,0.0070062221752272705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,7168,0.018045332696702745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,10240,0.03388177686267429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,4096,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,16384,0.022441777918073866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,65536,0.015111999379263984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,32,0.007585778004593319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,1024,0.013419555293189155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,512,0.008486222061845992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,128,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,768,0.015471110741297403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,32,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,4096,0.022467555271254644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,1536,0.03250577714708116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,256,0.0033697777738173804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,6144,0.0105608892109659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,4096,0.03581422236230638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,7168,0.041197333070966936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,12288,0.023006222314304773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,7168,8.234041849772135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,10240,0.06255555815166898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,3584,1.429981337653266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,16384,0.06781599919001262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,2048,0.011692444483439127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,2048,0.015522667103343539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,3584,0.1422942214541965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,65536,0.06398399670918782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,4096,0.0251217782497406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,512,0.004206222378545337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,4096,0.008340444829728868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,1024,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,5120,0.03867466582192315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,3584,0.08052533202701144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,4096,0.024634665913052026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,65536,0.012078222301271228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,65536,0.08689155843522813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,512,0.013395555317401886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,1536,0.006280888699822956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,3584,0.026595556073718604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,4096,0.007327111230956183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,7168,0.02769688930776384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,3584,0.03555644551912943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,2048,0.06212177541520861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,256,0.008351999852392409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,6144,0.018596443865034316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,32,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,7168,0.014448889427714877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,128,0.019647111495335896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,64,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,65536,0.05371910995907254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,64,0.0031964443624019623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,768,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,768,0.01718755563100179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,256,0.02199555602338579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,768,0.006136888845099344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,2048,0.01862844493654039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,12288,0.1195253398683336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,4096,0.008323555191357931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,256,0.02827022141880459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,4096,0.010039110978444418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,10240,0.09130311012268066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,768,0.0037715555065208008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,1536,0.016857778032620747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,5120,0.012411555482281579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,64,0.007053333024183909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,128,0.0033253334048721525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,1536,0.01761155492729611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,128,0.0028862222615215513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,10240,0.01681066718366411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,65536,0.763845337761773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,7168,0.17422844303978813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,2560,0.017607111069891188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,256,0.015173332558737861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,64,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,512,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,12288,0.018238221605618794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,256,0.029203553994496662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,12288,0.018577777677112155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,8192,0.020605333977275424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,2048,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,12288,0.07620000176959567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,5120,0.026392888691690233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,16384,0.13981866836547852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,10240,0.20018577575683594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,5120,0.016581333345837064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,1024,0.003590222034189436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,65536,2.204753875732422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,10240,0.2536826663547092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,7168,0.010018666585286459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,6144,0.026143999563323125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,5120,0.047833777136272855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,7168,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,1024,0.010047111246320937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,256,0.0034311111602518293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,1024,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,16384,0.045068442821502686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,128,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,128,0.008008889026112026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,64,0.0068959999415609576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,256,0.007620444728268518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,64,0.003522666792074839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,1024,0.13055288791656494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,64,0.005848000033034219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,65536,0.0737875567542182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,7168,0.01032977799574534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,128,0.026920000712076824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,65536,0.12740711371103922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,32,0.0038924444880750445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,12288,0.007679111427730984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,2560,0.009368889033794403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,5120,0.017086222767829895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,3584,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,2560,0.07171466615464953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,3584,0.028879112667507593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,2048,0.010597333312034607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,7168,0.013398222625255585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,2048,0.006619555668698416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,2560,0.006887110984987683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,1024,0.009489778015348647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,128,0.005199111170238919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,4096,0.00793333351612091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,3072,0.12150311470031738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,3584,0.007713778151406183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,1536,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,512,0.030396444929970637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,1536,0.006451555424266391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,256,0.005210666606823604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,1536,0.004858666823969947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,768,0.0038399998512532976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,1536,0.04224088788032532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,128,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,4096,0.01370133293999566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,2048,0.009672889278994666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,768,0.01612444387541877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,65536,0.05538666579458448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,1024,0.07720533344480726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,128,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,7168,0.021617778473430212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,3584,0.007612444460391998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,256,0.0032248888164758682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,128,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,10240,0.027845333019892376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,4096,0.02909066610866123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,7168,0.01036800030204985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,768,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,256,0.015793777174419828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,2560,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,768,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,12288,0.013400000002649097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,12288,0.007594666547245449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,2048,0.006923555499977536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,6144,0.059317330519358315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,6144,0.008977777428097194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,10240,0.022090666823916968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,256,0.01685244507259793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,7168,0.041535112592909075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,512,0.003896000070704354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,2048,0.006518222391605377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,2048,0.011583111352390714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,65536,0.2519386609395345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,256,0.01681600014368693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,10240,0.12972799936930338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,1536,0.012740444805887012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,512,0.007096889118353526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,3584,0.0069520001610120135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,1536,0.018477333916558158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,64,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,512,0.004143110993835661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,16384,0.01552888916598426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,2560,0.012390221986505719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,10240,0.012421333127551608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,6144,0.009023110899660323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,512,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,128,0.01477866702609592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,1536,0.0064088888466358185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,512,0.015819556183285184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,768,0.008030222521887885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,768,0.01811733345190684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,5120,0.010229333407349056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,4096,0.016002666619088914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,4096,0.007707555260923173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,1024,0.0057706667317284485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,768,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,6144,0.013255111045307584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,768,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,12288,0.014264000786675347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,2048,0.006704000135262807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,65536,0.0580328901608785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,128,0.0029271110478374693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,256,0.018294221825069852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,5120,0.021474666065639917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,10240,0.08656177918116252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,65536,0.6225048700968424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,256,0.0038942222793896994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,128,0.23138133684794107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,1536,0.006301333506902059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,10240,0.01276711126168569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,768,65536,0.09145955244700114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,12288,0.01864177816443973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,512,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,4096,0.007976888782448238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,7168,0.012208888928095499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,768,0.016612443659040663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,3072,0.09265955289204915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,10240,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,3584,0.11993243959214951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,7168,0.038935999075571694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,3584,0.01960000064637926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,16384,0.01885955532391866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,16384,0.024455111887719896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,65536,0.12956889470418295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,6144,0.008823110825485652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,10240,0.014488000008794995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,5120,0.012102222277058495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,10240,0.016516443755891588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,64,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,4096,0.010284444524182213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,2560,0.03568533394071791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,8192,0.05609155363506741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,4096,0.0148053334818946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,16384,0.01956888867749108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,512,0.012906666431162091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,6144,0.013246222502655454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,12288,0.0310444434483846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,1536,0.01484266585773892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,7168,0.033738666110568576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,4096,0.013713777893119387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,128,0.011396444506115384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,12288,0.03079644507831997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,2560,0.007628444168302748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,768,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,8192,0.035515556732813515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,65536,0.17065689298841688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,3072,0.011863110793961419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,128,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,12288,0.06308888726764254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,8192,0.018922666708628338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,6144,0.02232355541653103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,512,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,512,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,2560,0.06413422028223674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,6144,0.1457386679119534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,2048,0.006558222075303395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,128,0.015544889701737298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,65536,0.0591191119617886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,768,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,1024,0.015555555621782938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,128,0.0028453332682450614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,2048,0.014799111419253878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,4096,0.09335111247168647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,32,0.012658666405412885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,65536,0.09721510940127903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,65536,0.17501155535380045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,2048,0.014550222290886773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,7168,0.010030222435792288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,6144,0.012606222596433429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,3584,0.008633777499198914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,3584,0.007269333634111617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,8192,0.02464444438616435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,6144,0.02835644284884135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,16384,0.02093066606256697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,768,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,3072,0.09837777747048272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,512,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,2048,0.02035911050107744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,5120,0.010000000397364298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,256,0.01688088807794783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,16384,0.7639146910773383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,12288,0.040004445446862116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,10240,0.9300844404432508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,65536,0.07937688959969415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,2560,0.006752000086837345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,3584,0.00741511086622874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,6144,0.015244444211324057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,7168,0.02414488957987891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,2560,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,10240,0.1713564395904541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,3584,0.029940442906485662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,128,0.0148026661740409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,64,0.0026515554636716843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,3584,0.007620444728268518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,8192,0.010704889065689512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,1536,0.014094221923086377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,1024,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,2560,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,512,0.006735111276308696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,3072,0.016599110431141324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,3584,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,8192,0.01039644413524204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,128,0.014866666661368476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,10240,0.009031111167536842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,1024,0.0041964443193541635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,7168,0.006908444480763541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,8192,0.023669333921538457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,512,0.0031635556370019913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,12288,0.008632000121805403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,256,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,3072,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,256,0.018191999859280057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,12288,0.046148445871141225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,64,0.00250133333934678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,6144,0.009385777844323052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,768,0.009535999761687385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,8192,0.01630311045381758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,128,0.0032319999817344877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,256,0.0031662223239739737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,16384,0.051948444710837476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,512,0.003855111284388436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,2048,0.07185422049628364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,512,0.013864888913101621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,65536,0.12590489122602674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,8192,0.023690667417314317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,4096,0.015105777316623263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,3072,0.013008889224794177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,16384,0.056427558263142906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,65536,0.21414400471581352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,1536,0.03267200125588311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,1536,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,3584,0.09639200237062241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,2048,0.043040888177024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,4096,0.05067111055056254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,512,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,2560,0.016061334146393668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,2560,0.007254222200976477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,16384,0.04485511117511325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,256,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,16384,0.02884977724817064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,128,0.006887110984987683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,6144,0.01108000013563368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,2560,0.030572444200515747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,8192,0.013020444247457715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,1536,0.026094221406512793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,3072,0.007897777689827813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,2560,0.012746666868527731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,10240,0.10675022337171768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,10240,0.011406222151385413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,2560,0.017512889371977914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,2048,0.03135822216669718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,3584,0.007707555260923173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,1024,0.007666666474607255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,768,0.006226666685607698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,3584,0.02737866673204634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,10240,0.06447022491031222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,2560,0.009309333231714036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,7168,0.019632889164818656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,768,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,65536,0.1415662235683865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,16384,0.12486844592624241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,768,0.0069102222720781965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,512,0.020956445071432326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,1536,0.006529777828190062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,128,0.01165600038237042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,16384,0.022760889596409265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,4096,0.007970666719807519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,3072,0.017648888958825003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,1536,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,7168,0.012155555188655853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,6144,0.016195555528004963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,65536,0.40338754653930664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,12288,0.014442667365074158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,10240,0.045165333482954234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,1024,0.018904889623324077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,768,0.006051555689838197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,10240,0.06725955671734281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,16384,0.037979556454552546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,3072,0.0075777777367168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,64,0.002943111169669363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,1024,0.04510133465131124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,768,0.013181333740552267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,768,0.0034657776769664553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,5120,0.010985777609878115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,5120,0.00868444475862715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,8192,0.07325066460503472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,2048,0.00656088896923595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,12288,0.05529955691761441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,2560,0.030813333061006334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,5120,0.10461244318220352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,512,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,5120,0.04294044441647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,32,0.006660444455014334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,2560,0.011550222006109027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,5120,0.014743111199802823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,768,0.011095999843544431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,7168,0.02299733294381036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,8192,0.010481778118345471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,3584,0.01978400018480089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,6144,0.030220445659425523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,8192,0.09526489178339641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,12288,1.477225727505154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,10240,0.011378666592968835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,64,0.0032248888164758682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,3584,0.014907555447684394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,256,0.005224888937340842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,5120,0.020392888122134738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,2560,0.014516444669829475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,1024,0.016568889220555622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,768,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,3072,0.01591466698381636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,32,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,8192,0.02548622257179684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,4096,0.014811555544535318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,256,0.00490133340160052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,1536,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,768,0.008349333372380998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,2560,0.07282311386532254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,2560,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,64,0.004933333231343163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,65536,0.0684773325920105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,128,0.015127999915017022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,256,0.0031973332580592898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,10240,0.027454222242037456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,16384,0.01700266698996226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,3072,0.08191200097401936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,512,0.011882666912343768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,6144,0.010412444670995077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,768,0.003556444413132138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,768,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,4096,0.03387377659479777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,3584,0.016286222471131217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,1024,0.012221333053376941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,128,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,12288,0.023410666320059035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,512,0.0036959999965296853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,1024,0.15555732780032688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,10240,0.1544017791748047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,7168,0.016843555702103507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,12288,0.03583377930853102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,5120,0.007408888803588019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,5120,0.05464533302519056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,10240,0.01844888925552368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,7168,0.01899466746383243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,3584,0.023476445012622412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,16384,0.009434666898515489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,3584,0.02093155516518487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,8192,0.12090399530198838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,256,0.005333333379692501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,8192,0.0617413322130839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,10240,0.15620444880591497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,768,0.03725333346260919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,768,0.003919110943873723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,128,0.002871111035346985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,4096,0.10323999987708198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,64,0.005626666463083691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,7168,0.013459555804729462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,768,0.0035964445107513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,1024,0.021328889661365088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,1024,0.04235466652446323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,1024,0.25811645719740123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,7168,0.008632889224423302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,2560,0.009404444032245213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,64,0.014597333139843412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,2048,0.008588444855478074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,8192,0.026722666290071275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,6144,0.050034665399127536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,2560,0.023839111129442852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,5120,0.008348444269763099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,5120,0.020664888951513503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,10240,0.011781333221329583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,3072,0.007839110990365347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,512,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,3584,0.03142222099834018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,8192,0.01257511145538754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,65536,0.08464799986945258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,16384,0.013129777378506131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,8192,0.01698933376206292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,12288,0.23127288288540312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,12288,0.12060888608296712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,7168,0.01791911158296797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,2048,0.010588444769382477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,10240,0.02648177742958069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,12288,0.012464000119103326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,7168,0.013888888888888888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,256,0.008038221961922115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,2048,0.07275822427537706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,3584,0.12082578076256646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,2048,0.01981511049800449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,12288,0.020394666327370536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,512,0.010078222387366826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,10240,0.017666666044129264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,4096,0.01999022232161628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,7168,0.02414488957987891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,65536,0.16123199462890625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,6144,0.009360888765917884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,12288,0.029836446046829224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,10240,0.021350221501456365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,1536,0.01034311122364468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,12288,0.013400000002649097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,2048,0.011400000088744693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,128,0.00944088896115621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,512,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,7168,0.010451555252075195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,65536,0.22729423311021593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,7168,0.022859555151727464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,1024,0.006664888726340399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,128,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,768,0.015578667322794596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,4096,0.033596442805396184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,3072,0.00700444479783376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,2048,0.019268444842762418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,6144,0.014825777875052558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,768,0.015210666590266757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,1024,0.008728000024954477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,512,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,3072,0.018963555494944256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,768,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,3072,0.015573332707087198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,128,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,1536,0.007335999773608313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,3072,0.007711999946170383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,4096,0.008039111064540016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,768,0.02995733420054118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,12288,0.01737955543729994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,32,0.003576888806290097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,12288,0.020081778367360432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,32,0.0028808888875775864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,65536,0.09104800224304199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,7168,0.023464888334274292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,768,0.007336888876226213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,65536,0.019709333777427673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,3584,0.006611555400821898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,2560,0.09573600027296279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,2048,0.026802667313151892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,1536,0.010724444356229572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,16384,0.05490755703714159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,1024,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,12288,0.025511110822359722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,1024,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,3072,0.0063075555695427795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,128,0.005968889014588461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,128,0.014904888139830695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,1024,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,128,0.015536889433860779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,16384,0.4708399772644043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,1536,0.007247111035717859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,4096,0.014318222800890604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,12288,0.021376889612939622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,512,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,512,0.02755555510520935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,4096,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,128,0.015132443772421943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,65536,0.0810026658905877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,1024,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,65536,0.05438399977154202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,128,0.005642666584915585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,3072,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,12288,0.01314933349688848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,6144,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,768,0.018570666511853535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,6144,0.018391999933454726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,7168,0.009707555174827576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,6144,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,64,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,12288,0.04181333382924398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,16384,0.009728888670603434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,16384,0.1453662183549669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,6144,0.012490666574902005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,512,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,7168,0.007930667036109501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,16384,0.03856711255179511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,12288,0.22972800996568468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,7168,0.4179031054178874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,768,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,128,0.12493422296312119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,128,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,7168,0.009745777481132084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,512,0.024381332927280005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,3072,0.038909332619773015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,256,0.0029164445069101122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,10240,0.03360266817940606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,5120,0.05571022298600939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,65536,0.07923022243711683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,12288,0.03685777717166477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,1536,0.03642222285270691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,2560,0.0069884442620807225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,16384,0.06080355246861776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,16384,0.2557973331875271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,12288,0.05119466781616211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,256,0.015395555231306287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,768,0.01817599932352702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,1024,0.01464000013139513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,256,0.0037919998996787598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,768,0.018373333745532565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,8192,0.00700444479783376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,8192,0.016919111212094624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,16384,0.022788445154825848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,3584,0.007915555602974361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,4096,0.011639110743999481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,10240,0.010087110930018956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,10240,0.018624888526068795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,3584,0.016550223032633465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,1024,0.0063697778516345555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,12288,0.009071110851234859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,4096,0.015107555521859063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,10240,0.028519110547171697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,3072,0.015840889679061044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,2560,0.12000977993011475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,128,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,768,0.027447111076778833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,1024,0.06166399849785698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,6144,0.012475555141766867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,7168,0.0455297761493259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,256,0.012466666599114736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,65536,0.31301599078708225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,256,0.016940444707870483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,2560,0.006572444405820634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,2560,0.019313777486483257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,256,0.003858666867017746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,6144,0.04479022158516777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,512,0.015622221761279635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,1536,0.00490133340160052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,5120,0.009359111388524374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,1024,0.005532444351249271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,6144,0.02110222147570716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,2560,0.009008888569143083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,3584,0.01871466636657715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,32,0.0036026665733920205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,8192,0.020640000700950623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,3072,0.008999110923873054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,512,0.012507555385430654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,1536,0.004972444640265571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,128,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,3584,0.10243111186557347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,768,0.0059057776298787855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,512,0.007178666690985362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,768,0.022575111852751836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,65536,0.06458133459091187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,12288,0.012804444465372296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,1024,0.006685333533419504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,1024,0.005037333402368757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,256,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,7168,0.0310204459561242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,3584,0.010123555858929953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,5120,0.029112888707054987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,4096,0.17264178064134386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,64,0.003167111012670729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,1024,0.004260444392760594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,1536,0.004559999952713649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,2048,0.009973333113723332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,2560,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,1536,0.014764444695578681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,2560,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,1536,0.009402666654851701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,2560,0.05667733483844333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,2048,0.007696888513035244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,64,0.010010666317409938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,1024,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,10240,0.05617422527737088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,2560,0.027943111128277246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,6144,0.031118220753139917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,1536,0.00902844468752543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,2560,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,2048,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,7168,0.009425777528021071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,1024,0.019677332705921598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,768,0.00599911105301645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,7168,0.028935111231274072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,3072,0.023760888311598036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,3584,0.007739555504586961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,4096,0.01572177807490031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,2560,0.02824266751607259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,2048,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,512,0.009724444813198512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,256,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,64,0.002862222285734283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,5120,0.012499555945396423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,3072,0.013777777552604675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,8192,0.021165332860416834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,16384,0.019927110936906602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,2048,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,12288,0.16046222050984701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,8192,0.03229066729545593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,3072,0.01146755533085929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,1024,0.00664533343580034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,7168,0.010447111394670276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,16384,0.01519555515713162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,128,0.003367111086845398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,16384,0.10324888759189182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,128,0.015171556009186639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,6144,0.010218666659461128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,5120,0.013542222479979197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,8192,0.011315555208259158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,768,0.005767999837795894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,3584,0.03225333491961161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,128,0.030312889152103003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,1024,0.006095110956165526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,256,0.022014222211307947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,3584,0.03712622324625651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,5120,0.0090515555606948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,512,0.006230222268237009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,8192,0.01758577757411533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,64,0.002846222163902389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,4096,0.0254737784465154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,128,0.002921777880854077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,65536,0.019337778290112812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,64,0.002833777831660377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,2560,0.03727466530270047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,12288,0.039055999782350324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,6144,0.026603556341595117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,768,0.016551110479566786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,12288,0.008367111285527548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,7168,0.011779555843936073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,32,0.002834666727317704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,10240,0.011704000333944956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,128,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,2048,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,10240,0.01626488897535536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,4096,0.011023999916182624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,1536,0.004550222307443619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,5120,0.016528000434239704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,256,0.012446222205956778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,256,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,1536,0.013078221844302284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,12288,0.016366221838527255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,6144,0.009688888986905416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,256,0.009666666388511658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,1536,0.00685422205262714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,65536,0.1556533310148451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,128,0.00565422202150027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,65536,0.01516533394654592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,10240,0.012516444755925072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,3072,0.007923555870850882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,128,0.010072888599501716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,6144,0.012747555143303342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,8192,0.020617778102556866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,512,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,2560,0.006983111302057902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,12288,0.03031555480427212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,512,0.010027555955780877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,2048,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,8192,0.014737778239780001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,16384,0.05554666784074572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,12288,0.014111111561457316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,12288,0.06439377864201863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,6144,0.035671999057133995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,16384,0.021029333273569744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,64,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,64,0.0032302221904198327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,12288,0.023011555274327595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,4096,0.008010666403505538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,256,0.0038986665507157645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,5120,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,256,0.015896888242827523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,512,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,3584,0.12516978051927355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,2560,0.017145777742067974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,768,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,5120,0.014170666535695394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,64,0.007000000112586551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,2560,0.01756977703836229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,4096,0.023455111516846552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,64,0.0032106666929192017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,7168,0.010037333601050908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,4096,0.014236445228258768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,10240,0.03170311119821336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,1024,0.00618577789929178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,65536,1.9216177198621962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,8192,0.01833777791923947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,128,0.009791110952695211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,1536,0.0092284447617001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,768,0.03359466791152954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,7168,0.009775111244784461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,2560,0.015156444576051501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,10240,0.015626667274369132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,5120,0.008917333351241218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,65536,0.2592257658640544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,64,0.006175111151403851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,3584,0.00721866637468338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,128,0.006680000159475539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,32,0.003395555540919304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,4096,0.2908862166934543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,2560,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,4096,0.01683644453684489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,5120,0.017059556312031217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,4096,0.008691555923885768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,65536,0.1290071143044366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,2560,0.007099555598364935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,6144,0.009005332986513773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,2048,0.007082666787836287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,16384,0.02605688903066847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,256,0.019987555013762582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,3584,0.014374223020341663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,2560,0.00739822205570009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,3584,0.04050666756100125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,256,0.017684444785118103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,65536,0.48053598403930664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,3072,0.036320888333850436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,65536,0.47376712163289386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,512,0.00867022242810991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,8192,0.010850666297806634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,3584,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,1536,0.04365333252482944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,7168,0.02384977704948849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,1536,0.012665777570671506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,6144,0.01757866640885671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,4096,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,3072,0.08417155345280965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,2560,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,2048,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,2048,0.07318488756815593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,512,0.005643555687533484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,3072,0.015531554818153381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,3072,0.006026666611433029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,12288,0.013428444663683573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,65536,0.1403262217839559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,512,0.005983110931184556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,1536,0.004906666775544484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,7168,0.08504088719685872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,8192,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,64,0.006685333533419504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,32,0.00463733325401942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,65536,0.09104000197516547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,4096,0.007754666937722101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,16384,0.037372443411085345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,5120,0.027248889207839966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,65536,0.05655022462209066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,3584,0.020544000797801547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,1536,0.01349688900841607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,12288,0.12462577554914687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,32,0.003931555483076307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,65536,0.05772622426350912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,768,0.005631999837027655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,768,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,16384,0.03849066628350152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,65536,0.18998932838439941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,768,0.01665511065059238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,768,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,2560,0.03457422057787577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,512,0.011432888607184092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,5120,0.009034666750166152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,65536,0.1963182157940335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,2048,0.010713777608341642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,1536,0.004950222041871813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,1024,0.017844445175594754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,16384,0.027806222438812256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,1536,0.053676442967520825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,16384,0.025128000312381323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,2560,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,4096,0.013263999587959714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,256,0.012514666550689273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,7168,0.16763556003570557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,512,0.014987554815080432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,32,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,1536,0.016202666693263583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,65536,0.6128622161017524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,7168,0.03108888864517212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,7168,0.013960000541475085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,2048,0.016522667474216886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,3072,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,512,0.038338667816585965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,12288,0.006375999914275275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,16384,0.035749332772360906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,12288,0.02951999836497837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,12288,0.24040977160135904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,128,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,8192,0.00903911143541336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,1024,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,12288,0.01833066675398085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,3072,0.010831111007266574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,8192,0.018942221999168396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,12288,0.08645777569876777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,3072,0.008868444297048781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,4096,0.016558221644825406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,1024,0.019683554768562317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,12288,0.036757333411110774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,5120,0.009463999834325578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,1536,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,768,0.02873777680926853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,768,0.0053377776510185665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,65536,0.13337688975863987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,768,0.006036444256703059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,1536,0.016200888488027785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,32,0.0029315555261241067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,128,0.0038577777643998465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,2048,0.013741333451535968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,5120,0.013479111095269522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,6144,0.019359999232821994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,512,0.003552888830502828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,256,0.015551111764378019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,6144,0.00924622184700436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,8192,0.024975111087163288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,64,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,32,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,1024,0.010120888551076254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,64,0.004592000196377437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,16384,0.021016000045670405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,256,0.015185778339703878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,3072,0.007657777931955125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,8192,0.013619555367363824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,5120,0.008376888930797577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,5120,0.009900444083743626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,16384,0.027553778555658128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,512,0.012778666284349231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,1024,0.004930666751331753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,65536,0.14845866627163357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,2560,0.00665244460105896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,1536,0.006531555619504716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,5120,0.010055999788973067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,65536,0.06477422184414335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,12288,0.4714426464504666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,64,0.008648888932334052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,10240,0.027796443965699937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,768,0.02753955622514089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,3072,0.008382221890820397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,1536,0.02035911050107744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,4096,0.008513777620262569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,12288,0.01519733336236742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,64,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,3072,0.01588888963063558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,7168,0.010059555371602377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,128,0.015207111835479736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,8192,0.03335111008750068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,2048,0.024068444967269897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,12288,0.030934221214718286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,7168,0.3484657870398627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,256,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,1024,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,2048,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,6144,0.06513333320617676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,512,0.011763555308183035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,10240,0.012221333053376941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,768,0.02496355606449975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,768,0.005295111073387994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,8192,0.02405866649415758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,7168,0.031071109904183283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,5120,0.0074133334888352295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,2560,0.007048889166778988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,1024,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,2048,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,256,0.004952888935804367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,128,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,5120,0.009027555584907532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,2048,0.0053386667536364655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,4096,0.008014221986134848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,64,0.0032302221904198327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,7168,0.03132088979085287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,512,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,8192,0.012073777616024017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,7168,0.010024000373151567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,2048,0.013806222213639153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,3072,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,2048,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,32,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,16384,0.034994665119383074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,16384,0.027168888184759352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,32,0.0052391112678580815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,2048,0.015858666764365304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,16384,0.021179555190934077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,3584,0.028747555282380845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,2560,0.019291554888089497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,65536,0.07686133517159356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,3584,0.026188444760110643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,7168,0.010055999788973067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,12288,0.021931555536058214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,512,0.04774311184883118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,2048,0.030414223670959473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,12288,0.11983556217617458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,4096,0.007612444460391998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,3072,0.01928711103068458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,3584,0.018622222873899672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,10240,0.1305555502573649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,2048,0.017636444833543565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,1024,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,10240,0.015453333655993143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,6144,0.023418666587935552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,3072,0.018982221682866413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,10240,0.053620444403754346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,7168,0.014319111903508505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,7168,0.028388443920347426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,7168,0.08398933543099298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,3584,0.019949333535300363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,3584,0.020378667447302077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,1024,0.007698666718271043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,4096,0.3873804410298665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,1024,0.0064293332397937775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,16384,0.14157599873012966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,512,0.003572444534964032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,1024,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,3072,0.06333600150214301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,32,0.00674044465025266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,768,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,768,0.017177777157889474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,2048,0.014800889624489678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,8192,0.03971111112170749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,2048,0.006537777682145436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,2560,0.006983111302057902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,8192,0.02745955520206028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,12288,0.016693333784739178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,2048,0.0248355550898446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,2048,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,64,0.0032053333189752367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,8192,0.01882755590809716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,2048,0.01221777747074763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,512,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,12288,0.05647199683719211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,3072,0.015804444750150044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,32,0.0035706667436493766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,256,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,65536,0.10060889191097683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,12288,0.02853333287768894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,8192,0.031165334913465712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,5120,0.029956443442238703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,4096,0.008025777836640676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,65536,0.05616977479722765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,3584,0.0090515555606948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,768,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,3584,0.013553777502642738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,1024,0.014084445105658637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,128,0.014182221558358936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,4096,0.03359199894799127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,1536,0.013922666509946188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,4096,0.015214223000738354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,12288,0.013536888692114087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,4096,0.007096000015735626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,6144,0.55421691470676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,768,0.01218933363755544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,2048,0.009973333113723332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,1536,0.01314133322901196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,10240,0.021027555068333943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,2560,0.01962222158908844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,768,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,768,0.017349332571029663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,2560,0.021436444587177698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,8192,0.01035377797153261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,3584,0.017266665895779926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,1024,0.01752000053723653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,8192,0.04147911071777344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,256,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,1536,0.0070062221752272705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,5120,0.008352888955010308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,10240,0.03795999950832791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,5120,0.1360337734222412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,8192,0.1346693303849962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,5120,0.012341332932313284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,4096,0.008975110948085785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,10240,0.011798222031858234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,128,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,2560,0.006628444625271692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,10240,0.3807617823282878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,6144,0.014140443669425117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,3584,0.017454221844673157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,65536,0.09755644533369277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,256,0.021818665994538203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,2560,0.009415999882751042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,768,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,128,0.014144000079896716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,3072,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,3584,0.013554666605260638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,64,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,1536,0.012431110772821637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,128,0.01475911173555586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,8192,0.04882488979233635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,16384,0.026437333888477747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,65536,0.04697333441840278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,3584,0.009734222458468543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,6144,0.008967999782827165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,1024,0.010103111465771994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,6144,0.030650665362675984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,65536,0.37552444140116376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,768,0.00518844442235099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,3072,0.010392000277837118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,7168,0.9568790859646268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,256,0.006365333166387346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,7168,0.022202667262819078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,65536,0.23212711016337076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,7168,0.025220443805058796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,4096,0.01424888935354021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,1536,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,65536,0.1038702196545071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,512,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,7168,0.009390222529570261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,4096,0.016545777519543965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,1536,0.014019555515713163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,8192,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,12288,0.01425155500570933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,64,0.0032097777972618737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,65536,0.07218044333987765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,768,0.005665777872006099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,4096,0.009521777431170145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,512,0.005701333284378052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,1536,0.05429511268933614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,256,0.006500444478458828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,5120,0.015579554769727917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,1024,0.013907555076811047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,10240,0.016505777835845947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,512,0.0034764446318149567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,768,0.01699822147687276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,64,0.015855110353893705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,4096,0.090229332447052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,3072,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,2048,0.006983111302057902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,6144,0.023916444844669763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,768,0.0037120001183615792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,6144,0.09472266832987468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,2048,0.008748444418112436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,3072,0.00795555528667238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,1024,0.008621333373917473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,128,0.02251022226280636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,2048,0.011152889165613385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,16384,0.04220266805754768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,8192,0.04567199945449829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,6144,0.02019377715057797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,3072,0.03204355637232462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,12288,0.016531555189026725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,16384,0.2991911040412055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,1024,0.008350222474998897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,512,0.03121244576242235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,768,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,1536,0.009692444569534725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,7168,0.022637334134843614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,1536,0.007318221860461765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,3584,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,768,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,2560,0.014522666732470194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,32,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,12288,0.071633776028951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,4096,0.007963555554548899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,256,0.0031591111587153543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,2560,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,5120,0.10710044701894124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,5120,0.01625155574745602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,2048,0.011658666862381829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,10240,0.034359110726250544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,12288,0.01348533315791024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,768,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,3584,0.04012177719010247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,768,0.018182223041852314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,3584,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,512,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,1024,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,2048,0.006193777753247156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,32,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,6144,0.025244444608688354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,6144,0.01256444470749961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,2048,0.015806222955385845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,4096,0.01648888819747501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,768,0.01722400055991279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,256,0.013530666629473368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,10240,0.02572622232966953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,3584,0.012401777837011548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,768,0.0036657779581016968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,4096,0.029358221424950495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,1024,0.06184622314241198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,16384,0.18125243981679282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,8192,0.26923733287387425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,4096,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,1024,0.01775911119249132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,64,0.0032222221295038858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,2048,0.009959110783206092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,12288,0.0307262208726671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,7168,0.009305777649084726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,512,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,65536,0.0195795562532213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,128,0.008416888614495596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,256,0.007109333243634965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,5120,0.021287999219364587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,768,0.014106666048367819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,8192,0.1481742196612888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,65536,0.08381866746478611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,4096,0.025788444611761305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,1536,0.0063457778758472866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,512,0.014168888330459595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,5120,0.021672889590263367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,256,0.0028684443483750024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,10240,0.021965333157115515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,8192,0.021038222644064162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,5120,0.08023555411232842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,1536,0.013830222189426422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,3584,0.07776088847054376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,1024,0.0090506664580769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,4096,0.02316177719169193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,768,0.017892445127169292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,768,0.004228444563017951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,6144,0.01791200041770935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,7168,0.014179555906189812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,5120,0.020199111766285367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,10240,0.025946666797002155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,512,0.011432888607184092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,2048,0.005261333452330695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,768,0.020074667202101815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,5120,0.01607466737429301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,4096,0.03253600001335144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,5120,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,16384,0.029779556724760268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,2560,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,16384,0.022709333234363135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,2048,0.006877333339717653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,4096,0.05567200316323174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,10240,0.011743110915025076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,1536,0.028556442923016016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,128,0.014767110347747803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,2560,0.006726222319735422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,16384,0.08910044696595933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,2560,0.1971128914091322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,7168,0.028346667687098186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,64,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,8192,0.05908888578414917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,768,0.003863111138343811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,128,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,2048,0.0396106673611535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,1024,0.012379555238617791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,1024,0.014783110883500842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,7168,0.01825066738658481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,768,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,3584,0.01828266680240631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,1024,0.00423200014564726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,7168,0.03013777732849121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,3072,0.011414222419261932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,1024,0.013807999591032663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,256,0.004952888935804367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,65536,0.09171199798583984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,12288,0.019623999794324238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,8192,0.00831911133395301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,1536,0.07678311400943331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,65536,0.09650933080249363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,6144,0.015809777710172866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,2048,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,1024,0.00776622196038564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,3584,0.007660444411966536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,12288,0.013108444710572561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,65536,0.05819377634260389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,64,0.00942488925324546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,768,0.016965332958433364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,8192,0.018277333842383493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,128,0.003170666595300039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,8192,0.04638133446375529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,512,0.013431111143694984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,5120,0.007996444072988298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,65536,0.5004977650112575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,768,0.02196177840232849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,1024,0.007982222570313348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,128,0.015245333313941956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,6144,0.021352888809310064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,1536,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,64,0.004590222405062782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,32,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,3584,0.05939910809199015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,1024,0.015559110376569959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,16384,0.03275377882851495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,1024,0.024004444479942322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,65536,0.0989066693517897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,65536,0.020298666424221463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,3584,0.014841778410805596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,7168,0.012112889024946423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,256,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,768,0.007253333098358578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,256,0.003149333306484752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,6144,0.008959111240175035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,2560,0.009090666969617208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,64,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,512,0.011496889094511667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,32,0.008015111088752747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,256,0.00850755555762185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,256,0.005341333233647876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,3584,0.007501333124107785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,65536,0.014807111687130399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,768,0.005624889085690181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,6144,0.06563644276724921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,2048,0.02405244443151686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,1536,0.013948443863126965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,2048,0.0069724445541699724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,256,0.004209777961174647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,1024,0.012663111090660095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,65536,0.056418663925594754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,7168,0.010286221901575724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,7168,0.008263111114501953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,4096,0.016746666696336534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,32,0.0036915557252036203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,4096,0.01127022256453832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,256,0.014597333139843412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,1536,0.018613333503405254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,2560,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,128,0.021405332618289526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,5120,0.022077333596017625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,512,0.05620888868967692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,65536,0.14286133978101942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,2560,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,12288,0.02165422174665663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,768,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,1536,0.0046675557063685525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,768,0.003937777959638172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,16384,0.1655422184202406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,1536,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,16384,2.2613644070095487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,4096,0.01620444489849938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,12288,0.035197334157096014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,4096,0.14094755384657118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,1536,0.013258666627936892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,2560,0.015838222371207345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,256,0.005547555370463266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,1024,0.004271111140648524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,10240,0.017535110314687092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,3584,0.008519111408127679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,512,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,12288,0.024010666542583044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,128,0.015103111664454142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,1536,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,12288,0.492927975124783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,2560,0.01292799992693795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,6144,0.014053333136770459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,6144,0.009012444151772393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,2560,0.029882666137483384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,256,0.021372444099850122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,256,0.01716444392999013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,3072,0.015018666783968607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,256,0.011459555890825061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,5120,0.009710222482681274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,3072,0.022226666410764057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,64,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,3584,0.04198399848408169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,16384,0.37893954912821454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,256,0.0052391112678580815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,4096,0.020610666937298246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,128,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,768,0.006367110957702001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,1024,0.0063662222690052455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,6144,0.013643555343151093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,4096,0.008051555189821456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,768,0.009764444496896531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,128,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,2048,0.005965333431959152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,12288,0.011111111276679568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,65536,0.009328888522254096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,3584,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,10240,0.29964976840549046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,12288,0.02032977839310964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,2560,0.031110223796632554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,3072,0.025773333178626165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,128,0.015570667054918079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,2560,0.054752889606687755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,16384,0.020922667450375024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,6144,0.015747555428081088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,128,0.006816888848940532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,7168,0.03071111109521654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,12288,0.013544888959990608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,512,0.012695111334323883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,768,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,3072,0.010424888796276517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,16384,0.02151022189193302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,128,0.007688889073001013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,5120,0.1312826606962416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,256,0.03990044527583652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,10240,0.012135110795497894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,16384,0.16318399376339382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,1024,0.006027555714050929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,32,0.003257777748836411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,3584,0.008017777568764156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,4096,0.008192888564533657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,6144,0.010026666853162978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,32,0.007333333293596904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,768,0.01626844373014238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,256,0.005731555736727185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,4096,0.009394666386975182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,512,0.018574222922325134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,3072,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,16384,0.03237955437766181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,5120,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,512,0.005680888891220093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,768,0.01825777855184343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,64,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,1024,0.011552888486120435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,3584,0.012595555848545499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,5120,0.03943022092183431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,7168,0.04048622316784329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,7168,0.0233297778500451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,2048,0.014327999618318347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,512,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,16384,0.037167112032572426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,3584,0.011926222178671094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,2560,0.01689511040846507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,768,0.015609777635998197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,64,0.0028568889117903183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,6144,0.04101955559518602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,512,0.003510222252872255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,256,0.005976888868543837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,4096,0.01680266691578759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,10240,0.012802666260136498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,512,0.011738667057620155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,6144,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,8192,0.021345777644051447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,32,0.015857777661747403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,6144,0.01611822181277805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,32,0.0028817777832349143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,256,0.025075554847717285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,10240,0.041879110866122775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,8192,0.03230311142073737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,3072,0.017234666479958426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,256,0.024804444776640996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,32,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,8192,0.011651555697123209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,7168,0.1434275574154324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,4096,0.013781333135233985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,8192,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,1536,0.0282008879714542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,64,0.0053031109273433685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,512,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,65536,0.17563377486334908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,3584,0.007855999800893996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,12288,0.013251555462678274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,10240,0.02513866623242696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,512,0.0131217779384719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,2048,0.03997244437535604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,64,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,7168,0.018292443619834054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,1536,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,768,0.006269333263238271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,3072,0.07410666677686903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,65536,1.844686296251085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,512,0.011705777711338468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,16384,0.029637333419587877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,32,0.016593777471118502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,1536,0.0554622213045756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,768,0.01791111131509145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,8192,0.17009333769480386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,10240,0.03328888946109348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,256,0.005294222384691238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,2560,0.00871377769443724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,8192,0.010822222464614444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,5120,0.01313688854376475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,10240,0.021760000122918025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,16384,0.060496886571248375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,128,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,7168,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,4096,0.31978845596313477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,7168,0.018238221605618794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,3072,0.012030222349696688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,3072,0.008971555365456475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,8192,0.0465866658422682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,4096,0.012156444291273752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,12288,0.016879111528396606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,128,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,128,0.015486222174432544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,32,0.0038844446341196695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,768,0.005174222091833751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,256,0.015819556183285184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,8192,0.03741600116093954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,4096,0.01519377695189582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,7168,0.009762666291660732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,12288,0.026112889250119526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,3584,0.011955555942323474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,65536,0.09558400180604722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,4096,0.012792000340090858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,6144,0.040638221634758845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,2048,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,16384,0.05703555875354343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,16384,0.1096666653951009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,2048,0.013844444519943662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,16384,0.08023555411232842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,65536,0.07748622364468045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,2560,0.02277066641383701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,2560,0.07039999961853027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,1024,0.044334222873051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,10240,0.053120000494851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,1024,0.006624889042642381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,1536,0.005254222287072076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,768,0.008777778181764815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,768,0.003915555361244413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,256,0.012170666621790992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,256,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,10240,0.021368000242445204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,4096,0.02623733381430308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,768,0.005640888793600931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,12288,0.01680799987581041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,64,0.0028844444702068963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,4096,0.036630223194758095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,12288,0.1743475596110026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,3584,0.014535110857751636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,128,0.005688000056478713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,10240,0.012299555871221753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,7168,0.027489778068330552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,3584,0.01147377739350001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,8192,0.007125332951545715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,2560,0.01738933391041226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,256,0.0064693333374129395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,2048,0.007452444069915348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,5120,0.015996444556448195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,10240,0.009463999834325578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,256,0.003995555556482739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,128,0.006102222121424145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,2560,0.009296889106432596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,6144,0.009619555539555019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,2048,0.015124445160230001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,128,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,64,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,8192,0.025685333543353613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,1024,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,8192,0.01349511080318027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,1536,0.013519110778967539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,7168,0.03057866626315647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,8192,0.031069331698947485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,1024,0.013180444637934366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,1536,0.01869511107603709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,3072,0.008823110825485652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,12288,0.025555556019147236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,512,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,6144,0.009405333134863112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,3072,0.016535111599498324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,65536,0.01531555586391025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,6144,0.017860444055663217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,128,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,256,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,8192,0.09852355718612671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,1536,0.017266665895779926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,4096,0.02939200070169237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,768,0.007256000406212277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,32,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,2048,0.007710222568776872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,1536,0.006606222026877933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,512,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,256,0.01725955473052131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,768,0.015825778245925903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,12288,0.029465778006447688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,6144,0.01107377807299296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,128,0.005566222386227713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,3072,0.007704888780911763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,256,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,12288,0.17544355657365587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,768,0.013584000368913015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,10240,0.03311911225318909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,3584,0.00796088907453749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,128,0.00527022240890397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,4096,0.023159111539522808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,1024,0.018730666902330186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,3584,0.5162071122063531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,6144,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,16384,0.23125065697564018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,4096,0.03125600020090739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,3584,0.00908622228437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,3584,0.03307022319899665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,8192,0.01070844464831882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,128,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,65536,0.09987111224068536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,768,0.05760266383488973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,3072,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,5120,1024,0.02176977859603034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,65536,0.532642682393392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,2048,0.006573333508438534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,256,0.020607110526826646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,8192,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,12288,0.019959999455346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,2048,0.12354755401611328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,512,0.008414222134484185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,32,0.0027928888383838865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,3072,0.017474666237831116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,12288,0.04632355438338386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,1536,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,7168,0.02497244377930959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,8192,0.03997244437535604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,65536,0.1135537756813897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,5120,0.01236088905069563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,1024,0.006304889089531369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,3072,0.018625777628686693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,32,0.010100444157918295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,1024,0.019648889700571697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,65536,1.247844484117296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,65536,0.06770400206247966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,768,0.02399377855989668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,12288,0.02065333392884996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,7168,0.36911998854743105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,256,0.003552000141806073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,3072,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,12288,0.03340889016787211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,10240,0.16703555319044325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,768,0.01367911116944419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,16384,0.02195022172398037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,1536,0.015233778291278414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,10240,0.016552888684802584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,768,0.02516355613867442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,5120,0.012083555261294046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,2048,0.011083555718262991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,65536,0.20528888702392578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,3584,0.05595022439956665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,8192,0.12097244792514378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,7168,0.010950222611427307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,3584,0.015311999453438653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,65536,3584,0.1344453361299303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,4096,0.015855999456511605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,768,0.051289776961008705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,512,0.026847110854254827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,256,0.014129777749379476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,16384,0.03467377689149644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,3072,0.007316444483068254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,65536,0.1392515500386556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,5120,0.02755022214518653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,12288,0.017530666457282174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,64,0.007802666889296637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,1536,0.013435555828942193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,768,0.01682577696111467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,8192,0.05218133330345154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,4096,0.025584000680181716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,768,0.007904888855086433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,5120,0.025976889663272437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,32,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,2048,0.022882666852739122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,512,0.02204533252451155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,8192,0.009014222357008193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,512,0.0035182221068276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,8192,0.010747555229398938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,256,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,10240,0.015150222513410779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,12288,0.021402666966120403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,32,0.008295999632941352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,2560,0.018586667047606576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,16384,0.09873244497511123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,768,0.016701334052615695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,10240,0.2113351159625583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,12288,0.04239555531077915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,1536,0.011341333389282227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,3584,0.015491555134455362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,10240,0.47736178504096133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,6144,0.04716533422470093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,1536,0.0053431110249625305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,768,0.013619555367363824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,16384,0.04763111140992907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,1536,0.006311111152172089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,768,0.006640888750553131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,128,0.006078222145636876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,32,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,5120,0.017536888519922893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,3072,0.013697778185208639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,256,0.01590577761332194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,256,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,16384,0.03117155697610643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,2048,0.009881777895821465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,8192,0.017619555195172627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,512,0.016559110747443307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,7168,0.020049777295854356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,128,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,128,0.014549333188268872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,512,0.010065777434243096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,256,0.006319111006127463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,2560,0.013487111363146039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,512,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,6144,0.011147555377748279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,7168,0.010281778044170803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,2560,0.0070159998204973005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,12288,0.028552889823913574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,6144,0.021506667137145996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,6144,0.26371200879414874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,3072,0.027424888478385076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,2048,0.0068604445291890045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,2048,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,128,0.005285333428117964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,32,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,128,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,4096,0.007744000189834171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,768,0.017232000827789307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,5120,0.03082044588194953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,16384,0.02329866588115692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,2560,0.010449777874681683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,128,0.015231110983424716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,1024,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,3072,0.029114666912290785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,2560,0.014521777629852295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,7168,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,8192,0.015669332610236276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,7168,0.009752888646390703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,256,0.0033119999700122406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,3072,0.007669332954618666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,3072,0.04028622309366862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,128,0.0029084444459941653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,32,0.0031600000543726813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,5120,0.022289777795473736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,768,0.0075048887067370946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,64,0.0034346667428811393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,256,0.01723377737734053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,1536,0.06813066535525851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,256,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,32,0.0035822221802340615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,10240,0.0362337761455112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,1024,0.08182755443784925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,128,0.009323555562231276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,4096,0.014863999353514777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,8192,0.010788444843557147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,65536,0.16821600331200492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,256,0.005664000080691443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,32,0.006670222100284364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,768,0.003911111089918348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,65536,4.345439910888672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,7168,0.014171555638313293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,768,0.030844443374209937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,5120,0.04244533181190491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,8192,0.2651200029585096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,1536,0.007669332954618666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,32,0.004024888906213972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,2560,0.015209777487648858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,65536,0.470919132232666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,4096,0.016576000385814242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,65536,0.07484266493055555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,32,0.01724799970785777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,7168,0.012994666894276937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,5120,0.06183911032146878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,1536,0.004479111068778568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,2048,0.044425778918796115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,256,0.014070222775141398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,6144,0.009407111340098912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,4096,0.02833333280351427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,5120,0.02649333410792881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,1536,0.009524444739023844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,5120,0.08795377943250869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,2560,0.008042666647169325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,65536,0.17242577340867785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,1024,0.04181066817707486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,2560,0.013632000320487551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,10240,0.03419288992881775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,4096,0.00564533347884814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,3072,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,8192,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,128,0.003236444460021125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,128,0.011189333266682096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,3072,0.017161778277821012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,256,0.005283555636803309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,512,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,128,0.006639110959238476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,768,0.016283555163277518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,2048,0.015119999647140503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,32,0.009655999640623728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,1536,0.023702222439977858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,2048,0.0096142225795322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,512,0.05703022082646688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,12288,0.046216888560189136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,256,0.00554044461912579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,256,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,12288,0.0610391100247701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,12288,0.019529778096410964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,1024,0.004866666677925321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,3584,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,768,0.016873778568373788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,6144,0.05371288789643181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,16384,0.015461333923869662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,128,0.0032951111594835916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,16384,2.085484398735894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,4096,0.011991110940774282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,65536,0.8514702055189344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,8192,0.029802666770087347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,1536,0.03285333183076646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,1536,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,64,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,12288,0.060896886719597705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,6144,0.016128000285890367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,1024,0.01584533353646596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,1536,0.015784000356992085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,2560,0.009000889129108852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,512,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,768,0.005195555587609609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,1536,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,2560,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,1024,0.03338755501641168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,64,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,32,0.010016889207892949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,4096,0.011351111034552256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,8192,0.00982488857375251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,3584,0.016529776983790927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,65536,0.07747644186019897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,16384,0.02267733381854163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,256,0.008198222352398766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,1024,0.009300444689061906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,10240,0.011402666568756104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,6144,0.06316800249947442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,8192,0.013954665925767688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,1024,0.01516888870133294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,2560,0.07977510823143853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,3584,0.007686222592989604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,1536,0.014521777629852295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,8192,0.01073066641887029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,3072,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,7168,0.007687999970383114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,7168,0.014575999644067554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,65536,0.24890576468573677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,10240,0.04399111204677158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,7168,0.021391110287772283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,128,0.00832622249921163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,512,0.00563733321097162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,64,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,5120,0.017051556044154696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,64,0.007607999775144789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,12288,0.013244444297419654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,6144,0.031166222360399034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,2048,0.007339555356237624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,3072,0.013702222042613559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,8192,0.021340444684028625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,65536,2.6627137925889754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,3584,0.018904000520706177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,8192,0.03979466689957513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,16384,0.15851467185550266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,256,0.006737777756320105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,128,0.006017777654859755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,4096,0.007703999678293864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,8192,0.014929778046078153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,2560,0.007384888827800751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,1536,0.006335111127959357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,16384,0.01628266606065962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,7168,0.045159111420313515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,10240,0.06638755400975545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,6144,0.01763911048571269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,2048,0.032664888434939914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,1536,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,512,0.011418666276666852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,12288,0.010040889183680216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,5120,0.008991111483838823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,3072,0.03218400147226121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,2560,0.03234400020705329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,8192,0.014504889647165934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,7168,0.18026488357120088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,8192,0.592416869269477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,256,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,65536,0.08911910984251235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,6144,0.018598222070270114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,65536,0.05589333507749769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,128,0.0029306666304667792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,128,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,12288,0.017322666115230985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,128,0.012160888976520963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,2048,0.013800889253616333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,128,0.006773333168692059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,512,0.005623999983072281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,1024,0.00849511143234041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,4096,0.011067555182509951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,32,0.0032426667296224167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,8192,0.03190844588809543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,8192,0.2252497673034668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,768,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,128,0.00831911133395301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,1024,0.005927111125654644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,7168,0.06485778093338013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,8192,0.021904889080259535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,3584,0.022989332675933838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,512,0.005039111193683413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,16384,0.18052355448404947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,5120,0.010142222046852112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,768,0.0069982219073507525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,12288,0.05458755625618828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,12288,0.036325332191255354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,2560,0.006892444358931647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,6144,0.03263288736343384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,8192,0.022600000103314716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,64,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,128,0.0063475556671619415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,10240,0.013198222551080914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,7168,0.17355555958218047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,32,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,65536,0.08580355511771308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,3072,0.006974221931563483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,3584,0.01997422178586324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,256,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,512,0.0946248902214898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,64,0.01143555591503779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,10240,0.05267111129230923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,256,0.00630666646692488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,2048,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,2048,0.022343110707071092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,3584,0.01421333352724711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,512,0.005853333406978183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,512,0.020484444167878892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,2560,0.017515555024147034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,4096,0.007695111135641734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,2048,0.01405422223938836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,10240,0.03462933169470893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,3584,0.016858667135238647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,768,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,10240,0.031743112537595965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,8192,0.031086222993002996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,2048,0.06688977612389459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,256,0.003975999852021535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,10240,0.011791999969217511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,1536,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,7168,0.02238311039076911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,4096,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,16384,0.08354489008585612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,768,0.017280888226297166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,2048,0.006048888795905643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,3584,0.008825778133339351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,1024,0.008279999925030602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,5120,0.009023110899660323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,1024,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,128,0.016580444243219163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,128,0.012094222009181976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,3584,0.009725333087974125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,7168,0.010059555371602377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,128,0.010076444182131026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,32,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,12288,0.02295644415749444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,16384,0.020999999509917364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,256,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,7168,0.009516444471147325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,256,0.006314666734801398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,2048,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,256,0.0038533334930737815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,7168,128,0.01900888813866509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,2560,0.017549332645204332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,2048,0.008927110996511247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,8192,0.012111110819710625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,4096,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,32,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,1024,0.020635555187861126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,1536,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,16384,0.01110755569405026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,5120,0.011504000259770287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,5120,0.027144889036814373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,128,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,3072,0.013148444394270578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,5120,0.01831644442346361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,6144,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,65536,0.12967199749416775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,64,0.0028568889117903183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,2048,0.023815110325813293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,2048,0.011049778097205691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,4096,0.019221333993805777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,1024,0.018246221873495314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,7168,0.13833422131008571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,8192,0.020294222566816542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,768,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,3584,0.010956444674068026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,5120,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,16384,0.03932355509863959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,1536,0.007888889147175683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,64,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,3584,0.1094471083747016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,16384,0.01586577792962392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,8192,0.030438221163219873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,1024,0.0057608890864584185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,256,0.018590221802393597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,256,0.0058702222175068325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,65536,0.166648891237047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,16384,0.4118986659579807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,512,0.027943111128277246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,3584,0.011735111474990845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,128,0.0031928889867332247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,512,0.024699555502997503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,8192,0.03990133272276984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,8192,0.013630222115251752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,12288,0.0188719994492001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,1536,0.017868444323539734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,10240,0.024682665864626568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,12288,0.03888266616397434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,512,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,16384,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,512,0.005555555638339784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,12288,0.28167200088500977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,6144,0.1049777799182468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,32,0.0028417776856157514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,6144,0.010574222438865237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,1024,0.014210666219393412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,10240,0.011729777687125735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,256,0.014734221829308404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,1536,0.1244000064002143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,768,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,128,0.0151555554734336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,2560,0.0069955554273393415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,2048,0.02682311170630985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,128,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,16384,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,10240,0.00959644466638565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,3072,0.016307555966907077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,7168,0.016016888949606154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,16384,0.39509420924716526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,65536,0.027552889453040227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,16384,0.03455377618471781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,256,0.013442666994200813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,10240,0.019575110740131803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,65536,0.0748586654663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,32,0.007085333267847697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,12288,0.032425777779685125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,5120,0.008657777474986183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,2560,0.008595555192894405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,3584,0.0069715554515520734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,6144,0.017567111386193168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,1536,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,128,0.003020444470975134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,512,0.005853333406978183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,2048,0.007166222565703922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,6144,0.05838133229149712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,768,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,768,0.005691555639108022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,4096,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,256,0.003551111039188173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,3584,0.03856888744566176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,4096,0.010552888943089379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,2048,0.0074373334646224976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,65536,1.670779546101888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,2560,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,2048,0.013613333304723104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,1536,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,6144,0.037111110157436795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,65536,0.06273511383268568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,3072,0.013968888256284924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,1024,0.014166222678290473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,5120,0.012823999755912356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,32,0.004628444297446145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,8192,0.030163556337356567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,1536,0.004298666699065102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,5120,0.01699466672208574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,256,0.005684444473849402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,2560,0.039928889936871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,16384,0.0503475566705068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,64,0.0028560000161329904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,8192,0.017306667235162523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,3584,0.01181244436237547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,8192,0.03193155593342251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,128,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,1024,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,3584,0.006970666348934174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,768,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,768,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,3072,0.008043555749787224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,5120,0.00904266701804267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,12288,0.01798577772246467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,1536,0.00600888869828648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,768,0.013631111217869652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,2048,0.006659555352396435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,5120,0.1973439984851413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,768,0.005966222120655908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,128,0.012785777449607849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,6144,0.02753066685464647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,3072,0.05501066644986471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,12288,0.01832355558872223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,10240,1.1507733662923176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,6144,0.020280000236299302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,5120,0.008370666868156856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,4096,0.02234399980968899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,3072,0.239099555545383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,10240,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,128,0.019985778464211356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,5120,0.016883555385801528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,256,0.00904266701804267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,1024,0.017321777012613084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,8192,0.026846221751636926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,128,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,64,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,1536,0.015547555353906421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,3584,0.008031999899281396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,3072,0.02658133374320136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,12288,0.027446221974160936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,3072,0.015215999550289579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,2560,0.014926221635606555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,5120,0.011663999822404651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,8192,0.0953048865000407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,12288,0.01717155509524875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,32,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,65536,1.0237013498942058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,64,0.007346666521496243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,1536,0.008278222547637092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,8192,0.018343110879262287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,768,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,768,0.004280000097221798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,2048,0.8246044582790799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,16384,0.047015110651652016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,16384,0.033711112207836576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,512,0.008634666601816813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,3584,0.0162426663769616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,5120,0.026485333840052288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,65536,0.8776373333401151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,8192,0.015215999550289579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,1536,0.008902221918106079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,10240,0.023709333605236475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,2048,0.01626311077011956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,128,0.0038568890757030914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,2560,0.019855111837387085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,512,0.02479377720091078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,768,0.004211555752489301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,2560,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,2560,0.14419733153449163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,8192,0.02130311065249973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,512,0.005335111171007156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,64,0.006688888702127669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,4096,0.014992000328169929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,256,0.017210667332013447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,2560,0.0069955554273393415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,6144,0.01826400061448415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,2048,0.01963200006220076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,3072,0.007360888852013483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,32,0.002895999906791581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,512,0.015152888165579902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,16384,0.10068621900346543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,2560,0.006132444573773279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,128,0.00408266650305854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,16384,0.020136000381575692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,4096,0.01774133410718706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,12288,0.28828001022338867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,5120,0.018015111486117046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,3072,0.024805333879258897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,2560,0.007037333316273159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,32,0.005252444495757421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,5120,0.0127368892232577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,8192,0.010455110834704505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,3584,0.05191822184456719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,512,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,3584,0.06710133287641737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,2560,0.029529776838090684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,10240,0.017156443662113614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,512,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,12288,1.0707795884874132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,768,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,1024,0.004538666870858935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,16384,0.4471217791239421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,5120,0.01013777818944719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,128,0.00350577798154619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,4096,0.02828533450762431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,12288,0.021372444099850122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,2560,0.04311466548177931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,16384,0.10843644539515178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,4096,0.016886222693655226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,12288,0.01347555551264021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,32,0.0381351113319397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,5120,0.01038933296998342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,2560,0.007287999822033777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,32,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,2560,0.012525333298577202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,512,0.014142221874660917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,4096,0.05444799860318502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,768,0.003972444269392225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,2560,0.008047111332416534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,3584,0.018100443813535903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,12288,0.027449778384632532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,3584,0.022993778189023335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,512,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,1536,0.0225262227985594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,3072,0.007640000018808577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,5120,0.028856890069113836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,7168,0.07548977931340535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,10240,0.4432026545206706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,32,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,1024,0.008268444074524773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,12288,0.023948444260491267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,2560,0.006989333364698622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,7168,0.01038755559258991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,256,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,1024,0.006239111224810283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,2560,0.04997333221965366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,64,0.0039013334446483185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,32,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,7168,0.013541333377361298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,3584,0.09860533475875854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,65536,0.020833777056799993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,5120,0.008343111309740279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,10240,0.07565689086914062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,2048,0.008529778156015608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,256,0.011129777464601727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,10240,0.10736711157692803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,3072,0.011054221954610614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,2048,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,128,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,12288,0.01126844435930252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,256,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,128,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,65536,0.06700177987416585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,1024,0.01202488856183158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,32,0.0038720000949170855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,4096,0.015671999918089975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,2560,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,3072,0.007848888635635376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,512,0.008013333711359236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,2560,0.0075146663520071245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,6144,0.022325333621766832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,3584,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,32,5120,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,64,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,1024,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,4096,0.014907555447684394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,2560,0.012102222277058495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,256,0.003536888708670934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,1024,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,2560,0.037857777542538114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,8192,0.026467555099063452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,10240,0.011128889189826118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,65536,0.06998488638136122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,6144,0.013542222479979197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,5120,0.008113777471913232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,10240,0.13413066334194607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,7168,0.01940088967482249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,7168,0.009495110975371467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,3584,0.054231110546323985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,8192,2.3577857547336154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,6144,0.03372888763745626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,3072,0.012132444315486483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,512,0.008633777499198914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,1536,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,2048,0.017637333936161466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,4096,0.013159111142158508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,16384,0.29677867889404297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,6144,0.019370666808552213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,64,0.0033528889632887314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,1024,0.006474666711356904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,64,0.010144888526863521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,1024,0.022613333331214056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,16384,0.029262221521801416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,8192,0.023772444989946153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,3072,0.1537582212024265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,2048,0.006396444307433234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,6144,0.023386667172114056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,512,0.020596444606781006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,768,0.01313777764638265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,1536,0.02047911120785607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,32,0.003273777870668305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,1536,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,65536,0.4342542224460178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,3584,0.015594666202863058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,7168,0.007631111476156447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,8192,0.01757333344883389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,10240,0.021426666114065383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,10240,0.3592382272084554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,256,0.00833777752187517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,3072,0.007224000162548489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,1024,0.018964444597562153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,7168,0.02221244408024682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,2048,0.00664444433318244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,3072,0.016556445095274184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,1024,0.018046221799320646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,8192,0.022384888596004907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,8192,0.0986444420284695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,768,0.007721777591440413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,16384,0.02163733376397027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,7168,0.013366222381591797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,3584,0.019126221537590027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,12288,0.02883466747072008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,10240,0.02071822186311086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,512,0.004319111092223061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,512,0.005471111171775394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,7168,0.02628444466325972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,12288,0.06816177897983126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,128,0.005660444498062134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,7168,0.015083556373914083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,3584,0.020733333296246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,8192,0.008023111356629265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,7168,0.017940445078743827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,3584,0.01790577835506863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,64,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,3584,0.007701333198282454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,128,0.05539822247293261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,6144,0.02370044423474206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,3072,0.01568266749382019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,4096,0.026142223013771906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,128,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,1536,0.009176889227496253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,64,0.00527555536892679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,512,0.01719466679626041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,512,0.003338666632771492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,768,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,7168,0.01718488832314809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,6144,0.009016888837019602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,5120,0.028759999407662287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,12288,0.018940443793932598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,1024,0.028106666273540918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,16384,0.015816888875431485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,512,0.003903999924659729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,2048,0.006624889042642381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,6144,0.01891999940077464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,65536,0.11179378297593857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,12288,0.051282667451434664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,3072,0.009670221971140968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,3584,0.017982222967677645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,65536,0.2744986746046278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,1536,0.015110222829712762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,2560,0.01519466605451372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,1536,0.006073777874310811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,3072,0.008353778057628209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,7168,0.0120319997270902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,16384,0.0377786656220754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,2048,0.009131555755933126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,3072,0.007975999679830339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,2560,0.03350311186578538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,3072,0.010640888578361936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,7168,0.011428444749779172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,512,0.015967999895413715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,12288,0.09516444471147324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,32,0.003043555551105075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,1024,0.009161777794361115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,1536,0.020384889509942796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,12288,0.014279999666743808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,512,0.004376000000370873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,256,0.004016889052258598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,32,0.09468266699049209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,128,0.015026667051845126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,1024,0.010370666782061258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,4096,0.00812977800766627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,2048,0.008975110948085785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,4096,0.006080889039569431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,7168,0.01756000022093455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,1024,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,2560,0.0070168889231152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,512,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,256,0.008066666622956594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,16384,0.02203111184967889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,1536,0.017478222648302715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,8192,0.020091555184788175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,1536,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,256,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,3584,0.006986666884687211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,16384,0.039641777674357094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,32,0.0038808890514903595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,1024,0.005664000080691443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,12288,0.038035555018319026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,768,0.004893333133724001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,65536,0.09636088874604966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,5120,0.01761244402991401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,64,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,256,0.005292444593376583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,3072,0.028500444359249536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,7168,0.12654844919840494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,1024,0.006836444553401735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,256,0.005223111146026188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,6144,0.03253688746028476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,8192,0.01688888834582435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,2048,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,128,0.012808889150619507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,7168,0.01918577816751268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,5120,0.01383733335468504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,512,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,768,0.018202667435010273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,1536,0.009346666435400644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,512,0.013463999662134381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,2048,0.052710221873389355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,8192,0.09624533520804511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,12288,0.03301511208216349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,64,0.008051555189821456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,4096,0.017888888716697693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,128,0.015160888433456421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,768,0.0038968887594011095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,128,0.007606222397751278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,256,0.015206222732861837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,5120,0.008723555339707268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,6144,0.0387386679649353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,2048,0.023914666639433965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,128,0.03866577810711331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,3072,1.4614115820990667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,3072,0.016022221909628976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,128,0.002918222298224767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,256,0.0032311110860771606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,12288,0.05602400170432197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,16384,0.06524799929724799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,16384,0.052764445543289185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,256,0.014843554960356818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,6144,0.1222826639811198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,65536,0.12584800190395778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,16384,0.12000444200303818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,768,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,2560,0.05330222182803684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,4096,0.02532533307870229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,5120,0.016523554921150208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,2560,0.012758221891191272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,8192,0.013345777988433838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,6144,0.025606221622890894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,256,0.005640888793600931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,768,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,7168,0.023592000206311543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,5120,0.008743999732865227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,256,0.003531555748648114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,10240,0.012780444489585029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,3584,0.04380444354481167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,3584,0.010077333284748925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,3584,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,5120,0.009032888544930352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,5120,0.012816888590653738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,2560,0.0069511110583941145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,65536,0.09772800074683295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,3072,0.017909333109855652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,768,0.05032889048258463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,256,0.014845333165592618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,2048,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,1024,0.006024000131421619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,16384,0.03675377700063918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,2560,0.01109688894616233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,3584,0.03362311257256402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,256,0.03079644507831997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,2560,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,256,0.03642933236228095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,1024,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,128,0.0035751110149754416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,256,0.014935111006100973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,4096,0.06802311208513048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,7168,0.019817777805858187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,1536,0.01959733333852556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,10240,0.17988533443874785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,8192,0.010832000109884473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,10240,0.011863110793961419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,256,0.005616000129116907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,3584,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,512,0.01826400061448415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,2048,0.006969778074158563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,2560,0.007273777491516537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,768,0.021353777911927965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,5120,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,4096,0.09561333391401504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,2048,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,16384,0.026800000005298193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,3584,0.029112001260121662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,7168,0.019907555646366544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,768,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,4096,0.016574222180578444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,10240,0.03085333440038893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,10240,0.027081777652104694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,7168,0.01992266707950168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,8192,0.036995556619432234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,5120,0.01352355546421475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,512,0.003539555602603488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,6144,0.014190221826235453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,512,0.004223111189073986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,1536,0.007999111380841997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,64,0.005576000031497743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,1536,0.00657155571712388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,10240,0.05079466766781277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,16384,0.025241777300834656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,16384,0.00833777752187517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,2560,0.014864888456132678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,4096,0.03898488812976413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,6144,0.01831999917825063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,1536,0.004544888933499654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,2048,0.008614222208658854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,16384,0.03811644514401754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,2048,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,1024,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,12288,0.020632000433074105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,3072,0.0148026661740409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,8192,0.016026667422718473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,1536,0.013589333328935834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,3072,0.01519466605451372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,256,0.005240889059172736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,16384,0.015631111131774057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,768,0.01943466729587979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,64,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,4096,0.3792577849494086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,1536,0.013148444394270578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,256,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,65536,0.17247645060221353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,256,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,16384,0.030066665675905015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,3584,0.01039644413524204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,4096,0.008077333370844523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,2560,0.36752088864644367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,12288,0.01724444495307075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,10240,0.09819200303819443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,768,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,10240,0.024365334047211543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,128,0.005990222096443176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,7168,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,3072,0.03547911180390252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,12288,0.06883822547064887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,768,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,2560,0.01574577722284529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,65536,4096,2.072286181979709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,128,0.026478222674793665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,2048,0.010675555302037133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,1536,0.018559111489189994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,512,0.02247200078434414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,16384,0.0524142218960656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,10240,0.03816533419820998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,65536,2.160519070095486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,128,0.013487111363146039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,6144,0.1130711105134752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,128,0.00433688900536961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,256,0.015933333171738517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,16384,0.019596444235907663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,12288,0.06985866361194186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,128,0.00886311133702596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,512,0.003489777859714296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,3584,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,2560,0.07642222113079496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,2560,0.012416889270146688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,1536,0.006355555521117316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,32,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,3584,0.07451111078262329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,3584,0.015643555257055495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,10240,0.1697404384613037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,12288,0.02314044369591607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,256,0.013888888888888888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,4096,0.01110311100880305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,256,0.0052533331844541765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,7168,0.0125591109196345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,3584,0.05852711200714111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,3072,0.0993644462691413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,128,0.006934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,6144,0.011732444167137146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,7168,512,0.00629333323902554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,1536,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,1536,0.05850044223997328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,10240,0.024674667252434626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,2048,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,2048,0.06470133198632134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,2048,0.008974221845467886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,5120,0.08165688647164239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,64,0.005723555468850666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,7168,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,65536,0.1998008886973063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,3584,0.007616889145639207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,2560,0.01238666640387641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,6144,0.05223200056287977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,128,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,65536,0.19017244709862602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,2560,0.012728888955381183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,16384,0.20739821592966715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,2560,0.0074782222509384155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,6144,0.01295377810796102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,1024,0.07785955402586195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,512,0.0075048887067370946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,3072,0.007778666913509369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,32,0.006747555401590135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,8192,0.01092888911565145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,5120,0.014330666926172046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,128,0.014956444501876831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,4096,0.012895111408498553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,128,0.014817777607176038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,65536,0.064173334174686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,7168,0.05041689011785719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,1024,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,1536,0.009386666946940953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,128,0.017983999517228868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,16384,0.019555555449591745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,128,0.011525332927703857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,256,0.016840888394249808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,2560,0.02202933364444309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,3584,0.007672888537247975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,32,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,4096,0.02901600135697259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,5120,0.011360000405046674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,3072,0.011755555868148804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,6144,0.10623288816875881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,768,0.00351200004418691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,32,0.013923555612564087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,7168,0.008601778083377415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,1536,0.023676445086797077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,16384,0.022806222240130108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,12288,0.009714666340086196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,512,0.05606310897403293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,65536,0.1425644424226549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,1024,0.0240320000383589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,4096,0.008042666647169325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,2048,0.05749066670735677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,10240,0.012087999946541257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,256,0.014547554983033074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,3072,0.037610666619406805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,5120,0.03163377775086297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,256,0.01833599971400367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,16384,0.015934222274356417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,1024,0.014200000299347771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,5120,0.2505013412899441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,5120,0.02762311034732395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,5120,0.015878222054905362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,2560,0.006989333364698622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,256,0.006014222072230444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,2048,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,6144,0.02096533278624217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,65536,0.04099199838108487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,1024,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,7168,0.0281315545241038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,7168,0.0196142229768965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,1024,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,10240,0.024068444967269897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,256,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,2048,0.0063031112982167145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,128,0.0032257777121331957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,10240,0.029176000091764662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,7168,0.16200621922810873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,16384,0.027448000179396734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,2048,0.011495999991893768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,6144,0.5313448905944824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,4096,0.011613333390818702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,768,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,768,0.043062223328484424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,128,7168,0.021386666430367365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,12288,0.02107288936773936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,12288,0.029998222986857098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,128,0.013139555851618448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,2048,0.02973066767056783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,32,0.004217777815130022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,1536,0.004576888763242298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,512,0.0035048888789282907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,128,0.011383111278216044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,3584,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,8192,0.010739555789364709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,12288,0.038113776180479265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,10240,0.03520177801450094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,3072,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,16384,0.029603554142846003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,6144,0.011104889214038849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,5120,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,65536,0.07988622453477648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,6144,0.007369777394665613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,2048,0.009020444419648912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,2048,0.01089599976936976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,2048,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,2560,0.007288888924651676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,16384,0.02372444503837162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,64,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,8192,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,7168,0.021661332911915247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,10240,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,8192,0.013557333085272046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,1536,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,64,0.0028417776856157514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,1024,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,12288,0.02548177705870734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,5120,0.020999999509917364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,3584,0.016561778055297006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,4096,0.010051555931568146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,2560,0.19232000244988334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,256,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,6144,0.013921777407328287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,8192,0.013935999737845527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,6144,0.2298053370581733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,5120,0.017814222309324477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,5120,0.02645422187116411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,1536,0.006326222171386083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,4096,0.026664889521068994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,2560,0.016415999995337594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,8192,0.046726223495271474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,1536,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,768,0.017234666479958426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,4096,0.006993778049945831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,3072,0.006620444357395172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,32,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,2560,0.00867377801073922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,5120,0.011714666253990598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,10240,0.025768000218603346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,10240,0.051905777719285756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,32,0.0035546666218174826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,128,0.008189332981904348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,2048,0.014827556080288358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,12288,0.03569333420859443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,512,0.01181422256761127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,16384,0.011391999820868174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,6144,0.051671110921435885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,1536,0.004592000196377437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,5120,0.018100443813535903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,16384,0.024056888288921777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,16384,0.060990220970577665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,256,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,3072,0.017535110314687092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,512,0.015533333023389181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,65536,0.06536444690492418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,7168,0.009407999614874521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,1024,0.008368888662921058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,4096,0.01461866663561927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,2560,0.012263110942310758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,5120,0.012816000315878125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,8192,0.09819200303819443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,6144,0.031574222776624895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,256,0.031098667118284438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,1024,0.015978667471143935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,768,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,12288,0.02612000041537815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,7168,0.012475555141766867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,6144,0.01961955593691932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,7168,0.02052088909678989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,7168,0.02609511050913069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,3072,0.03782222337192959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,12288,0.018519111805491976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,32,0.010381333529949188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,512,0.0135057775510682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,12288,0.04725689027044508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,2048,0.05372800098525154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,3072,0.0103502223889033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,2560,0.011847111086050669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,2048,0.006983111302057902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,16384,0.010112000008424124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,8192,0.15841155582004124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,256,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,1536,0.006247111078765657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,6144,0.039373334911134504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,256,0.013798221945762634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,256,0.015108444624476962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,512,0.011400889191362592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,8192,0.2011200057135688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,1024,0.009663999908500248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,10240,0.24000265863206652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,512,0.0032239999208185407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,128,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,65536,0.10394489102893406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,2560,0.012154666913880242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,512,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,12288,0.026452443665928308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,1536,0.02349600030316247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,1024,0.006270222365856171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,1024,0.037252446015675865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,2048,0.024499555428822834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,3584,0.008009778128729926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,512,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,2048,0.04632888899909126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,12288,0.013463111387358772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,8192,0.011767999993430244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,12288,0.01313777764638265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,65536,0.15022755993737116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,16384,0.021011556188265484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,3584,0.27491733762953013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,5120,0.00888355573018392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,768,0.023608888188997906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,1536,0.11840621630350749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,3584,0.021740444832377966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,1024,0.0052915554907586836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,3584,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,128,0.02201333310869005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,10240,0.007335999773608313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,256,0.015541333291265698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,65536,0.5658515294392904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,64,0.003935111065705617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,128,0.004592888885074192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,512,0.022749332918061152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,3072,0.008028444316652086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,1536,0.005976888868543837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,2048,0.013204444613721637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,7168,0.03602755400869582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,768,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,4096,0.007672000262472365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,16384,0.09460533327526516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,768,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,32,0.006608888920810487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,8192,0.07341155740949842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,3584,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,1536,0.004746666798988978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,12288,0.013451555536852943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,7168,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,8192,0.021827555365032617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,2560,0.03205422229237027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,3584,0.05574755536185371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,1024,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,1024,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,64,0.002842666581273079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,8192,0.007257777783605788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,5120,0.015603555573357476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,768,0.004113777644104428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,32,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,768,0.003826666623353958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,3072,0.01586844523747762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,3072,0.007537778053018782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,12288,0.010384000009960597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,768,0.01824711097611321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,6144,0.013649777405791812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,6144,0.017440888616773818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,2560,0.02195555633968777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,3072,0.012741333080662621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,1536,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,256,0.002903999967707528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,16384,0.026351110802756414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,12288,0.017177777157889474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,16384,0.5899510913425022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,1536,0.0090515555606948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,65536,0.07320444451438056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,3072,0.01758044461409251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,3072,0.010753778119881948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,2560,0.011885333392355176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,2048,0.007069333559936947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,4096,0.01441511180665758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,4096,0.015602666470739575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,65536,0.14616178141699895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,768,0.03412177827623155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,10240,0.03210399879349603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,2560,0.05563644568125407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,1024,0.003912888881233004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,4096,0.007632888853549957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,1024,0.0058106668293476105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,6144,0.009318222602208456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,1536,0.007956444389290279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,512,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,16384,0.0151582227812873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,1536,0.008608889248636033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,1024,0.01181422256761127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,1536,0.007785778078767989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,5120,0.00870577742656072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,3072,0.007613333563009898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,512,0.0032506665835777917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,3072,0.015358222855461968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,16384,0.017627555463049147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,10240,0.03840622305870056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,3584,0.14470577239990234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,2048,0.008049777812427944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,4096,0.06029866801367866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,3584,0.018552889426549275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,3584,0.0251226673523585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,512,0.0558533337381151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1024,3072,0.015159999330838522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,8192,0.01959377858373854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,128,0.005584888988071018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,2048,0.009078222016493479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,2560,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,2560,0.018923555811246235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,12288,0.022209778428077698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,65536,0.06061866548326281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,5120,0.017236444685194228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,16384,0.05883555279837715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,5120,0.019109333554903667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,32,0.010041777458455827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,6144,0.020564445190959506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,6144,0.018235555953449674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,3584,0.011397333608733283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,64,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,512,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,2560,0.006879111131032308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,1024,0.015124445160230001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,4096,0.02291733357641432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,5120,0.0484497778945499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,768,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,1536,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,256,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,3072,0.013568888935777875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,3072,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,2048,0.007375111182530721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,32,0.01108888867828581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,16384,0.020290666156344943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,4096,0.007970666719807519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,12288,0.01759022143152025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,8192,0.023965333898862202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,512,0.0572835538122389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,256,0.0034808889031410217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,7168,0.014129777749379476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,2048,0.0696444445186191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,8192,0.01571733256181081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,5120,0.025423111187087163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,2048,0.021703999903466966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,7168,0.02648711038960351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,7168,0.010011555420027839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,16384,0.00656088896923595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,1024,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,6144,0.009387555221716562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,2560,0.006986666884687211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,16384,0.030890666776233252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,12288,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,10240,5120,0.30011378394232857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,5120,0.007256000406212277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,3584,0.016499555773205228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,512,0.0075039996041191956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,1024,0.007617777420414819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,5120,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,768,0.005646222167544895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,7168,0.02681333323319753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,10240,8192,0.04268444577852885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,32,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,10240,0.012831110921170978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,3072,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,256,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,12288,0.015358222855461968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,2560,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,512,0.014450665977266101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,1024,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,5120,1.3128560384114583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,16384,0.00943199959066179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,512,0.007330666813585494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,2048,0.07869599925147162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,7168,0.010472888747851053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,2560,0.024104000793562994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,64,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,10240,0.03930933276812235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,12288,0.06026577949523926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,8192,0.18050221602121988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,3584,0.017235555582576327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,4096,0.020434666011068556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,3072,0.015338665909237333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,8192,0.017312000195185345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,7168,0.02165688905451033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,2048,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,2048,0.005223111146026188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,2560,0.02754488918516371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,10240,0.18254133065541586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,512,0.0047066667013698155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,7168,0.023583999938435022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,8192,0.03411733441882663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,16384,0.025123554799291823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,4096,0.03657599952485826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,32,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,4096,0.01183555523554484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,6144,0.01182577759027481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,1536,0.015800888339678448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,8192,0.03475022315979004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,2048,0.025217778152889673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,12288,0.013228444589508904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,12288,0.017438222964604694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,12288,0.27546665403578013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,8192,0.010888888604111142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,2048,0.0063893331421746155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,1536,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,6144,0.007989333735571967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,3584,0.01650311052799225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,128,0.007631999750932057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,768,0.007289778027269576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,5120,0.017072889539930556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,256,0.006239999913507038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,8192,0.20917867289649117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,16384,0.03950577643182542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,5120,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,5120,0.009723555710580613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,2048,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,12288,0.050344000260035195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,16384,0.03740444448259141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,10240,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,12288,0.23006667031182182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,7168,0.02029511166943444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,2048,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,10240,0.0692195561197069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,6144,0.024117334021462336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,1536,0.008282666405042013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,3584,0.010024000373151567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,256,0.00350577798154619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,5120,0.009685333404276106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,128,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,12288,0.014083556003040738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,128,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,7168,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,10240,0.030128889613681372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,12288,0.19410578409830728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,16384,0.038210666841930814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,12288,0.016506666938463848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,768,0.015814221567577787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,8192,0.010375111467308467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,2048,0.0052382221652401825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,8192,0.026704889204767015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,8192,0.15887378321753606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,12288,0.0329733325375451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,8192,0.03932888971434699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,512,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,4096,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,2048,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,1536,0.004979555391603046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,512,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,64,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,8192,0.040088001224729754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,4096,0.007631999750932057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,65536,0.1882035599814521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,1024,0.005969777703285217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,512,0.020433778564135235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,4096,0.055848002433776855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,2560,0.015563555889659457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,1024,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,16384,0.024688889582951862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,7168,0.04633333285649618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,3072,0.015832000308566622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,2048,0.017571555243598092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,5120,0.01756000022093455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,65536,0.2541617817348904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,1536,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,1024,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,768,0.005903999838564131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,128,0.002887111157178879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,1024,0.006225777582989798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,64,0.0034675554682811103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,7168,0.032518221272362605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,5120,0.020011555817392137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,10240,0.04059911105367873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,64,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,1536,0.01587466730011834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,10240,0.03681333197487725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,1536,0.018192888961897958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,1024,0.009334222310119206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,256,0.015497777197096082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,65536,0.11947288778093125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,65536,0.07903911007775201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,5120,0.016194666425387066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,2048,0.020572443803151447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,2048,0.02424266603257921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,10240,0.02438844409253862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,2048,0.01549600064754486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,8192,0.08682133091820611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,768,0.0058702222175068325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,512,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,10240,0.016418667303191293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,16384,0.03156355685657925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,7168,0.022711111439598933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,1536,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,10240,0.011074666347768573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,3072,0.017825777331988018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,1024,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,1536,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,1536,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,1024,0.007932444413503012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,3584,0.007621333003044128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,256,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,8192,0.025623111261261836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,1536,0.00675733346078131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,256,0.005694222119119432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,2560,0.04391910963588291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,512,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,6144,0.028208888239330713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,65536,0.16748355494605172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,10240,0.009751111268997192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,1536,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,1024,0.31662845611572266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,1024,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,256,0.0035111109415690103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,128,0.01789955629242791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,3072,0.00830844458606508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,512,0.19617422421773276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,768,0.008839999636014303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,8192,0.16526133484310573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,2560,0.008589333130253686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,8192,0.011145778000354767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,65536,0.08061955372492473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,8192,0.11180355813768174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,8192,0.04558044340875414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,6144,0.15214488241407606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,2560,0.01037599974208408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,10240,0.012080888781282636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,768,0.014793778459231058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,2560,0.016857778032620747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,16384,0.03513955407672458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,65536,0.17504444387223986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,16384,0.04275555411974589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,3072,0.013076444466908773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,7168,0.03396000133620368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,1536,0.015983111328548856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,32,0.019623999794324238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,2048,0.015801777442296345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,7168,0.016235555211702984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,8192,0.020641777250501845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,7168,0.1628097825580173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,128,0.014144889182514615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,12288,0.026473777161704168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,2560,0.03630933496687148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,2560,0.006599999964237213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,1024,0.007642666498819987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,8192,0.04208177659246656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,128,0.0052782222628593445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,1024,0.024665777881940205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,256,0.003391999958289994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,10240,0.02682755556371477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,128,0.011377777490350934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,4096,0.011559111376603445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,128,0.0028906667398081887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,10240,0.04527555571662056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,768,4096,0.016149333781666227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,6144,0.030107554462220933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,1024,0.008267555799749162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,12288,0.09883644183476765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,1024,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,2048,0.004939555707905027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,3072,0.021015110943052504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,65536,0.07648089196946886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,4096,0.012671999633312225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,64,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,256,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,7168,0.01238666640387641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,7168,0.1280871099895901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,128,0.015488889482286243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,7168,0.013809777796268463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,16384,0.016184889607959323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,1024,0.007756444315115611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,3584,0.015672889020707872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,65536,0.01533777846230401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,16384,0.13413422637515598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,128,0.015280000037617154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,1024,0.12433511681026882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,3584,0.02647200061215295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,65536,0.08762133121490479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,5120,0.010019555687904358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,512,0.011376889215575324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,8192,0.030089779032601252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,65536,0.5598782433403863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,512,0.008650666309727563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,8192,0.03481333454449972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,16384,0.04876088764932421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,8192,0.01484266585773892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,3584,0.01072799993885888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,3072,0.019497777024904888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,1024,0.008291555775536431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,12288,0.024716445141368445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,10240,0.03653511073854234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,3072,0.007602666815121968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,16384,0.219777782758077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,16384,0.01682400041156345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,1024,0.0038373333712418876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,64,0.0028702221396896574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,128,0.015543111496501498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,256,0.003863111138343811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,1024,0.00600355573826366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,7168,0.21072178416781953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,7168,0.3193342155880398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,7168,0.21542489528656006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,6144,0.03201155530081855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,512,0.006689777804745569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,32,0.004226666771703296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,512,0.011759111450778114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,5120,0.053123556905322604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,1024,0.01347644461525811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,1024,0.009681777821646797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,4096,0.034635557068718806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,12288,0.015799999237060547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,2560,0.015445333388116626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,64,0.003002666764789157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,7168,0.02474755545457204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,5120,0.011075555450386472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,65536,1.378061294555664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,2048,0.058416002326541476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,2560,0.012261333564917246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,2048,0.009578666753239101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,10240,0.10240978002548218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,65536,0.2595706780751546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,5120,0.022486221459176805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,128,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,12288,0.0236453331179089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,256,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,7168,0.0295688874191708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,512,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,65536,0.07515288723839654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,10240,0.012729778057999082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,65536,1.2541369332207573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,128,0.0031902222997612427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,12288,0.017771555317772757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,256,0.00527022240890397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,7168,0.009721777505344814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,1536,0.004581333448489507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,64,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,512,0.014651555154058667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,128,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,8192,0.11480177773369683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,1024,0.006082666830884085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,1536,0.010011555420027839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,4096,0.017208000024159748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,512,0.014674666855070325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,10240,0.2589120070139567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,2560,0.014155555102560254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,1536,0.027078222897317674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,65536,0.055454221036699086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,3584,0.02221511138810052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,2048,0.01349244432316886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,512,0.013815110756291283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,32,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,3584,0.01682577696111467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,10240,0.04258666767014397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,3584,0.007962666451931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,32,0.03734133309788174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,8192,0.022487110561794702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,2048,0.016985777351591323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,3584,0.009494221872753566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,128,0.020615110794703167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,128,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,64,0.0028151110228565005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,2560,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,10240,0.024811555941899616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,7168,0.009317333499590555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,5120,0.041944887903001576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,3072,0.023008000519540574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,2048,0.024911110599835713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,4096,0.017437333861986797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,7168,0.010107555323176915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,3584,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,2560,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,4096,0.011358222199810876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,1024,0.0041759999261962045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,768,0.003921777837806278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,32,0.009684444301658207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,12288,0.032848887973361544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,16384,0.020229332976871066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,10240,0.0155404441886478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,7168,0.03527822097142538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,12288,0.02074399921629164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,2560,0.005868444426192178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,64,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,12288,0.018295110927687753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,128,0.015490666031837463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,8192,0.010738666686746808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,1536,0.028934223784340754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,6144,0.02052800026204851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,1024,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,64,0.0025759999536805684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,12288,0.048820445934931435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,3584,0.021777777208222285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,7168,0.042282667424943715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,2560,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,1536,0.01037333326207267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,1536,0.019944889677895438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,3072,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,1536,0.09154044257269965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,65536,0.3091119925181071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,10240,0.05426399906476339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,1536,0.019280888968043856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,2560,0.0070017774899800616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,7168,0.030555556217829388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,768,0.003867555409669876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,1536,0.01644533375898997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,4096,0.006069333189063602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,16384,0.020785777105225455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,64,0.01012711144155926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,12288,0.05200533403290642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,5120,0.05650222301483154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,8192,0.010791999598344168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,3584,0.010311110979980892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,256,0.017154667112562392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,16384,0.04686399963166979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,65536,0.06189688709047106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,1536,0.004587555511130227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,16384,0.01109777804878023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,7168,0.013200889031092325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,5120,0.010375111467308467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,1536,0.004936888813972473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,512,0.0032115555885765287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,768,0.007889777421951294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,12288,0.01826400061448415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,256,0.010784000158309937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,1024,0.003976888954639435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,5120,0.016275554895401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,2048,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,3072,0.016538666354285348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,1536,0.004942222187916438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,10240,0.04680266645219591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,32,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,2048,0.008015111088752747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,2048,0.006956444846259222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,1024,0.004503111044565837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,4096,0.019464888506465487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,768,0.003726222034957674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,512,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,64,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,1536,0.007983999947706858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,768,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,768,0.004602666530344221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,12288,0.05119199885262383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,2048,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,10240,0.022675555613305833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,64,0.0041742221348815495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,3072,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,1536,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,128,0.0053004444473319584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,16384,0.01972888906796773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,512,0.016509332590632968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,256,0.03391822179158529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,3584,0.04953866534762912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,1536,0.015102222561836243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,1024,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,3072,0.031896889209747314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,128,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,3072,0.00904799997806549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,128,0.0059057776298787855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,256,0.00566133318675889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,4096,0.030720889568328857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,2560,0.01053955571519004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,65536,0.36644710434807676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,16384,0.04068266683154636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,256,0.005530666559934616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,128,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,64,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,3072,0.011047110789351992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,16384,0.020929776959949072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,1024,0.014476444986131458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,256,0.005191111316283544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,7168,0.00980266680320104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,65536,0.304879109064738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,8192,0.029564443561765883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,7168,0.017900443739361234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,512,0.005792888916201062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,10240,0.041668444871902466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,3072,0.008623110751310984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,2560,0.0105591110057301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,10240,0.011749332977665795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,32,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,12288,0.016951999730534024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,65536,0.061964445643954806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,6144,0.012095111111799875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,12288,0.022967111733224656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,2048,0.013291555146376291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,1024,0.016206221448050607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,1024,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,5120,0.010082667072614035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,5120,0.020918221937285528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,3072,0.010377777947319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,128,0.0031191110610961914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,12288,0.06464444266425239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,7168,0.019840000404251944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,512,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,1536,0.04004177782270644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,3584,0.020447111792034574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,64,0.008024888734022776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,10240,0.012054222325483957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,7168,0.02602488961484697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,10240,0.03206222256024679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,12288,3.7568435668945312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,512,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,7168,0.019926221834288705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,6144,0.0924817787276374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,12288,0.0365226666132609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,16384,0.13093688752916124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,65536,0.2060533364613851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,2048,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,2048,0.09445866611268784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,768,0.06462755468156603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,16384,5.4664416842990455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,64,0.01071911139620675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,4096,0.014112000664075216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,3072,0.012094222009181976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,10240,0.023172444767422144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,64,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,10240,0.0570088889863756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,7168,0.17395199669731987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,256,0.016206221448050607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,6144,0.02680088910791609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,2560,0.02989333205752903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,8192,0.016521778371598985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,65536,0.07117333014806111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,128,0.01548266741964552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,128,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,10240,0.018030222919252183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,7168,0.0158933334880405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,256,0.008263111114501953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,3584,0.011750222080283694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,10240,0.04196177919705709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,16384,0.022499556342760723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,65536,0.1423964500427246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,1536,0.010383110907342698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,4096,0.3016435570187039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,256,0.006672888994216919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,7168,0.01462222304609087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,16384,0.822347535027398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,1024,0.013417777915795645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,2560,0.019567999574873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,16384,0.04447911183039347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,16384,0.05832978089650472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,12288,0.049588445160124034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,3584,0.19821066326565215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,128,0.008743999732865227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,16384,0.03436888919936286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,8192,0.08060710959964328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,6144,0.014476444986131458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,512,0.00833155545923445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,256,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,2048,0.006586666736337874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,7168,0.010110222631030613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,512,0.008049777812427944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,8192,0.031268444326188825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,128,0.015146666102939181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,1024,0.014126222994592456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,768,0.004555555681387584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,512,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,2560,0.0058017778727743365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,16384,2.113532384236654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,3072,0.08521333005693223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,10240,0.014793778459231058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,1536,0.006606222026877933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,6144,0.010496888723638324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,768,0.017247110605239868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,1024,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,1536,0.00832177781396442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,128,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,1024,0.017281777328915067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,65536,0.14645244015587702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,10240,0.026081777281231348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,768,0.017283555534150865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,2048,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,768,0.005741333381997214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,5120,0.007970666719807519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,3072,0.017308443784713745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,128,0.005196444276306364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,32,0.008280889027648503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,1024,0.015456888410780164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,3584,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,256,0.016530666086408827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,6144,0.013003555436929068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,512,0.024919110867712233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,1536,0.009716444545321995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,1536,0.027691556347741023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,128,0.015455111861228943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,256,0.015142222245534262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,8192,0.05500444438722399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,3584,0.008256889051861232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,5120,0.00867911097076204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,5120,0.018255111243989732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,10240,0.006211555666393704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,3584,0.008349333372380998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,512,0.005573333137565189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,5120,0.008332444561852349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,2560,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,4096,0.012085333466529846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,512,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,256,0.00700444479783376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,6144,0.01696977847152286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,3584,0.03831111060248481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,5120,0.053527110152774386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,1024,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,3072,0.008237332933478886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,10240,0.05005066593488058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,16384,0.09176088704003228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,3584,0.012574222352769641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,256,0.01719555589887831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,2048,0.006764444626039929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,1536,0.009166222479608323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,4096,0.05851022402445475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,1536,0.02753511071205139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,65536,0.062000890572865806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,16384,0.05722577704323662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,16384,0.04491200049718221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,1536,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,128,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,768,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,1536,0.006721777634488211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,1536,0.024097777075237695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,2560,0.01259022206068039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,10240,0.014197332991494073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,7168,0.20114044348398843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,16384,0.020592000749376085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,2048,0.03705244594150119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,65536,0.09001333183712429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,8192,0.025399999486075506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,256,0.003362666608558761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,768,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,3072,0.006745777610275481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,6144,0.03679822219742669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,2048,0.008667555948098501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,512,0.013730666703648038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,4096,0.027496889233589172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,8192,0.02269777821169959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,1024,0.004965333475006951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,3584,0.016943110360039603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,12288,0.021260443660948012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,3584,0.009352000223265754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,12288,0.11488178041246201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,3072,0.006490666833188798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,3072,0.007356444166766272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,5120,0.10970755418141682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,16384,0.023949333363109167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,16384,0.03923111160596212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,4096,0.018025777406162687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,10240,0.016578666037983365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,65536,0.13013688723246256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,512,0.019646222392718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,512,0.005487111293607288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,512,0.007974222302436829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,3584,0.01901155544651879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,10240,0.028004444307751123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,10240,0.013092444174819522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,7168,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,256,0.016857778032620747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,2048,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,10240,0.027241778042581346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,3072,0.007617777420414819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,3584,0.014416888356208801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,64,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,65536,0.06226666768391927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,256,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,7168,0.011771555576059552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,16384,0.035304887427224055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,10240,0.0836168858740065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,512,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,256,0.006874666859706243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,10240,0.01479111115137736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,12288,0.08084711101320055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,3584,0.022942221826977197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,12288,0.02496444516711765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,16384,0.13198310799068874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,2048,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,768,0.005541333307822545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,128,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,2560,0.01127911110719045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,12288,0.12132799625396729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,1536,0.033497777250077986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,4096,0.008956444760163626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,12288,0.08201422293980916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,128,0.010017777482668558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,12288,0.03199733297030131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,768,0.007548444800906711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,8192,0.01588355501492818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,2560,0.01037688884470198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,10240,512,0.010375111467308467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,5120,0.011692444483439127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,4096,0.024878222081396315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,10240,0.029193778832753498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,4096,0.010047111246320937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,4096,0.007298666569921706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,12288,0.06426666842566596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,5120,0.020595555504163105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,12288,0.014140443669425117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,10240,0.22144622272915312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,2560,0.006640000061856375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,768,0.013421333498424955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,64,0.0028577778074476454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,12288,0.031183110343085393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,1536,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,16384,0.0359226663907369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,6144,0.020471110939979553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,512,0.016532444291644625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,1024,0.006119110931952794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,7168,0.01073333372672399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,5120,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,128,0.0036115555299652945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,64,0.0028533333291610083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,5120,0.008385777473449707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,10240,0.049831112225850425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,2560,0.012372444073359171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,7168,0.128602663675944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,10240,0.014684443672498068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,10240,0.015464888678656684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,2048,0.009316444396972656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,256,0.0033804445217053094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,6144,0.019567999574873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,12288,0.03154133425818549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,3072,0.042731556627485484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,768,0.016880000631014507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,1024,0.004377777791685528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,6144,0.008995555341243744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,768,0.030537777476840552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,1024,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,128,0.009405333134863112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,768,0.025472889343897503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,6144,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,6144,0.009370666411187913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,8192,0.037921776374181114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,10240,0.01716088917520311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,3584,0.01833777791923947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,7168,0.04363377888997396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,4096,0.02462311089038849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,3584,0.02252799934811062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,10240,0.03054311209254795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,3584,0.008298666940795051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,2048,0.006928000185224745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,256,0.0031902222997612427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,2560,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,512,0.006555555595291986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,256,0.012436444560686747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,7168,0.012472888661755456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,12288,0.019208888212839764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,3584,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,3072,0.022053332792388067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,1024,0.00629333323902554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,5120,0.016157334049542744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,6144,0.013757333159446716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,512,0.021288888321982488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,16384,0.023656888140572443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,768,0.013385777672131857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,5120,0.026802667313151892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,512,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,32,0.0032133333798911837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,3584,0.015837333268589444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,2048,0.006480888773997624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,4096,0.015480889214409722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,2048,0.014352889524565803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,256,0.003548444559176763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,16384,0.06756444772084554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,2048,0.011423110961914062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,2048,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,1536,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,256,0.00314666661951277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,768,0.007309333317809635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,65536,0.05899288919236925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,128,0.02053866618209415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,6144,0.041137778096728854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,2048,0.006911110960774952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,65536,0.17608711454603407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,12288,0.03260355525546604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,65536,0.136409772766961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,128,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,65536,0.09765511088901097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,4096,0.017875555488798354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,1024,0.006630222416586346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,2048,0.034535109996795654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,10240,0.029071112473805744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,512,0.01479822231663598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,32,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,6144,0.02404622236887614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,8192,0.030869334936141968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,7168,0.008524444368150499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,128,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,65536,0.15666133827633327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,6144,0.011063111325105032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,2048,0.0148026661740409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,65536,0.04970577690336439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,16384,0.016229333149062265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,7168,0.01719466679626041
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,12288,0.019293333093325298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,64,0.009767110976907942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,32,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,3072,0.01752622259987725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,2048,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,512,0.003843555433882607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,5120,0.024517332514127094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,12288,0.2410337660047743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,2048,0.00699733363257514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,10240,0.058281779289245605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,2560,0.012434666355450949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,5120,0.008020444048775567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,4096,0.0614888866742452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,65536,0.07432177993986341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,5120,0.015004444453451367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,256,0.00924800005224016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,12288,0.013478221992651621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,2048,0.010590222146775989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,768,0.00665155549844106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,10240,0.018608000543382432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,16384,0.04061066773202684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,10240,0.029863112502627905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,5120,0.05177066723505656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,8192,64,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,10240,0.046804444657431714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,1024,0.018976888722843595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,1536,0.007913777397738563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,128,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,6144,0.008716444174448649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,8192,0.03379911184310913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,6144,0.058281779289245605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,4096,256,0.03306044472588433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,256,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,1024,0.015142222245534262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,2048,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,12288,0.030505776405334473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,3072,0.04998488889800178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,10240,0.015564444992277356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,1536,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,768,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,2048,0.009955555200576782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,256,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,7168,0.04152889053026835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,768,0.013847110999955071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,64,0.002883555574549569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,6144,0.0682613319820828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,32,0.002830222249031067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,4096,6144,0.011123555401961008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,2048,0.007611555357774098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,5120,0.007930667036109501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,1536,0.00452622233165635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,256,0.02346844474474589
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,65536,0.08275110854042901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,3584,0.011052444577217102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,4096,0.008332444561852349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,8192,0.06210044357511732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,10240,0.023364444573720295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,7168,0.00997866690158844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,2048,0.009248889154858058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,6144,0.10360178020265366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,3584,0.380021333694458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,7168,0.02978488802909851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,3584,0.007991111112965478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,1536,0.004520888957712385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,10240,0.023358222511079576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,10240,0.012316444681750404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,6144,0.01401422255569034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,512,0.06549155712127686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,5120,0.013411555853154926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,1536,0.060256891780429415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,512,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,1536,0.01199288914601008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,7168,0.030826667944590252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,512,0.00599644457300504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,512,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,512,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,6144,0.020291555258962844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,2560,0.0929715567164951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,768,0.003548444559176763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,7168,0.02612888813018799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,768,0.016947555873129103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,128,0.007640000018808577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,16384,0.008006222546100616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,8192,0.010665777656767102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,256,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,3584,0.007615110940403408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,7168,0.032482667101754084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,5120,0.009348444640636444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,5120,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,256,0.004202666795916027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,12288,0.02788177794880337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,512,0.00867733359336853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,4096,0.08896355496512519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,10240,0.03773066732618544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,2048,0.015199111567603217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,32,0.0028977776981062363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,5120,0.16408356030782065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,65536,0.064246224032508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,5120,0.01256444470749961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,65536,0.19972977373335096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,2048,0.032371554109785296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,8192,0.0185208883550432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,1536,0.07728888591130574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,1536,0.08164355489942762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,256,0.006565333240562015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,2048,0.005592888842026393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,4096,0.007916444705592262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,10240,0.009280888570679558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,7168,0.07485066519843207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,7168,0.02055199940999349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,5120,0.10811822281943427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,16384,0.16018311182657877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,10240,0.053058667315377124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,4096,0.009538667069541084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,4096,0.008719111482302347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,2560,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,10240,0.01650311052799225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,768,0.02141600019401974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,1536,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,3072,0.016536000702116225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,4096,0.03289955523278978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,3072,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,3584,0.08705955743789673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,1536,0.004604444321658877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,6144,0.018191999859280057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,3584,0.016487111647923786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,64,0.005528000079923206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,256,0.016504888733228047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,6144,0.018195556269751657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,64,0.022082666556040447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,3072,0.006955555743641323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,6144,0.031552000178231135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,7168,0.020047111643685233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,1024,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,4096,0.010853333605660332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,32,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,1536,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,3072,0.007994666695594788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,10240,0.02331377731429206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,3584,0.020988444487253826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,8192,0.010682666467295753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,10240,0.011738667057620155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,32,0.0028248888750871024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,1024,0.009693333672152625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,3072,0.06829777691099378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,3072,0.008302222523424361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,8192,0.007232888705200619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,8192,0.026126222478018865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,5120,0.049103998475604586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,2048,0.004912888838185204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,1536,0.02647999922434489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,3072,0.007332444190979004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,4096,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,5120,0.01789066692193349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,128,0.014870221416155497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,5120,0.3168133364783393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,16384,0.013651555611027611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,128,0.016151999433835346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,64,0.008052444292439355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,4096,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,12288,0.04188977678616842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,2560,0.00702400008837382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,65536,0.061978664663102895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,7168,0.020733333296246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,10240,0.015208888385030957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,3584,0.03327111072010464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,8192,0.010399111443095738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,65536,0.07716177569495307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,128,0.014811555544535318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,8192,0.01571377780702379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,7168,0.026557332939571802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,10240,0.0576488905482822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,256,0.011027555498811932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,2560,0.02715822226471371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,8192,0.022610666023360357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,1024,0.004206222378545337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,6144,0.014515555567211576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,12288,0.026106667187478807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,64,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,256,0.0203164451652103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,1024,0.006262222097979651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,2560,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,1536,0.08793689144982232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,2560,0.019665777683258057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,2560,0.01072266697883606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,1024,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,64,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,8192,0.012086222569147745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,1024,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,768,0.006198222024573221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,3072,0.007251555720965068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,512,0.004920888692140579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,4096,0.011071110765139261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,2560,0.014803555276658801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,64,0.01620533400111728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,128,0.029623111089070637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,5120,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,64,0.005656888915432825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,768,0.01755466726091173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,3584,0.011352000137170156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,256,0.01648533344268799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,8192,0.04869066675504049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,128,0.0031519999934567344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,6144,0.12807822227478027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,12288,0.03192889028125339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,8192,0.04821688930193583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,4096,0.008005333443482717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,3584,0.017870222528775532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,512,0.006224000205596288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,512,0.0052426668504873914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,256,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,8192,0.489033752017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,32,0.005714666512277391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,3584,0.008046222229798635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,8192,0.008022222254011366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,10240,0.03536266750759549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,768,0.01744888888465034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,4096,0.01166311071978675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,8192,0.027015111512607996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,128,0.003927111211750242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,4096,0.01587288909488254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,4096,0.04460888769891527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,2560,0.03145600027508206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,128,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,16384,0.01811200049188402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,65536,0.022798221972253587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,7168,0.012903999951150684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,16384,0.015612444943851896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,12288,0.013204444613721637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,6144,0.02012266715367635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,3072,0.015544000599119397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,6144,0.00999022192425198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,768,0.0035439998739295532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,256,0.014965333872371249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,512,0.013720889058378009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,8192,0.012732444538010491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,256,0.017482666505707633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,8192,0.02239111065864563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,7168,0.01204177737236023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,2048,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,3072,0.026672000686327618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,32,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,256,0.005208888815508948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,16384,0.05003022154172262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,1536,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,1536,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,1536,0.022621333599090576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,512,0.056454221407572426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,65536,0.06919288635253906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,12288,0.015226667126019796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,512,0.01923288901646932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,128,0.006224000205596288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1536,1024,0.004208000169859992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,1536,0.01202400028705597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,65536,0.220104005601671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,65536,0.09106400277879502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,3584,0.02857422166400486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,3072,0.013806222213639153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,4096,0.02199999988079071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,10240,0.02290311124589708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,3584,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,32,0.011374221907721626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,1024,0.007694222033023834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,2048,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,2048,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,1024,0.15016267034742567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,1536,0.005503111001518037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,10240,0.013740444348918067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,32,0.0052737775776121355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,4096,0.056121779812706836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,768,0.004861333303981357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,3584,0.008738666772842407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,512,0.006193777753247156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,2560,0.0258542216486401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,16384,0.010320000350475311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,512,0.0052968888647026485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,65536,0.1101377805074056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,128,0.02628622286849552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,128,0.006960888703664144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,10240,16384,0.08667377630869548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,65536,0.137463993496365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,1536,0.014488000008794995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,5120,0.021027555068333943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,16384,0.020103999310069613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,1024,0.054618666569391884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,16384,0.04522844486766391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,12288,0.017314665847354464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,1536,0.010986666712496014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,3072,0.007261333366235097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,1024,0.062430222829182945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,1536,0.022076444493399724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,6144,0.017838221457269456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,5120,0.027740443746248882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,3072,0.022844443718592327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,12288,0.06683378087149726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,128,0.0028026666906144884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,4096,0.012980444563759698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,1536,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,1536,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,1536,0.012440889245933957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,2048,0.006573333508438534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,5120,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,2048,0.00797333319981893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,7168,0.02401777770784166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,3072,0.010029333333174387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,5120,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,8192,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,6144,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,64,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,10240,0.02826933397187127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,6144,0.023191110955344305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,65536,0.10104533036549886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,2560,0.005818666683302985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,12288,0.02722399930159251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,3072,0.009660444325870937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,3072,0.036299556493759155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,768,0.006766222417354584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,6144,0.026174222429593403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,12288,9.878273010253906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,5120,0.026165333059098985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,7168,0.03267822331852383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,768,0.010023111270533668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,3072,0.02834133307139079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,6144,0.035968889792760216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,65536,0.15253599484761557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,512,0.010794666906197866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,5120,0.024785778588718835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,2048,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,32,0.003112888791494899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,3072,0.007290666302045186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,512,0.003839111162556542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,128,0.005633777628342311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,6144,0.0075048887067370946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,1536,0.007684444387753804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,3584,0.010434666441546546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,2048,0.012144889268610211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,128,0.0028817777832349143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,7168,0.18250577979617646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,1024,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,12288,0.03646399908595615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,1536,0.00518133325709237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,128,0.0029173334025674393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,768,0.006302222195598815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,6144,0.018726223044925265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,4096,0.012101333174440594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,3072,0.007664889097213745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,12288,0.05316444569163852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,6144,0.018201778332392376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,2560,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,6144,0.007954667011896769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,16384,0.08540444241629706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,256,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,512,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,1536,0.005201777650250329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,1024,0.0048942222363419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,3072,0.09165333376990424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,2048,0.006243555496136348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,512,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,128,0.009703111317422655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,3584,0.018287110659811232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,7168,0.01719555589887831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,256,0.015812445018026564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,512,0.01369866645998425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,16384,0.020996444755130347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,1024,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,128,0.006897777732875612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,1024,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,128,0.0328106681505839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,8192,0.013423110875818463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,5120,0.014304889572991265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,4096,0.019640889432695176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,2560,0.009903999666372934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,2560,0.010683555569913654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,2048,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,256,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,128,0.01775644388463762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,128,0.014514666464593677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,768,0.004574222283230888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,8192,0.02033955521053738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,6144,0.023502222365803186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,256,0.015120888749758402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,3072,0.01629333363638984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,4096,0.19984977775149873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,16384,0.22012355592515734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,16384,0.019573332534896005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,1536,0.01514933341079288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,8192,0.01406222250726488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,12288,0.037371555964152016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,1024,0.02973066767056783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,3072,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,512,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,6144,0.023593778411547344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,2048,0.006667555620272954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,4096,0.00795555528667238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,1024,0.027648000253571406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,3072,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,4096,0.008295111358165741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,128,0.015140444040298462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,128,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,16384,0.2401342127058241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,10240,2.5636649661593967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,3072,0.022192888789706763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,128,256,0.01552177800072564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,256,0.01611111064751943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,12288,0.0295413335164388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,512,0.02237066626548767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,8192,0.021330666210916307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,2560,0.015789333317014907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,32,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,5120,0.1081813308927748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,128,0.0032871110985676446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,5120,0.018038221531444125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,64,0.003067555526892344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,7168,0.05075644453366598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,64,0.003661333272854487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,2560,0.01389244364367591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,128,0.015221332510312399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,10240,0.02287111183007558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,65536,0.033682667546802096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,10240,0.026728888352711994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,7168,0.019686222076416016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,128,0.006010666489601135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,1536,0.00600622221827507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,512,0.006683555742104848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,512,0.02167911165290409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,3072,0.007369777394665613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,512,0.003624888757864634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,768,0.003567999849716822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,512,0.057619558440314404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,1536,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,3072,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,5120,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,3584,0.17864533265431723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,3584,0.026032888227038916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,10240,0.021669333179791767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,128,0.011988444460762871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,10240,0.017904000149832833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,1536,0.015488889482286243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,1024,0.054231110546323985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,65536,0.09146310885747273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,2560,0.0784791111946106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,6144,0.020346666375796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,7168,0.013228444589508904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,5120,0.020392888122134738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,256,0.0031519999934567344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,256,0.025919111238585577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,65536,0.07598933246400622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,4096,0.010337777435779572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,256,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,32,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,4096,0.02164266672399309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,10240,0.015533333023389181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,256,0.01310755560795466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,2048,0.013083555632167392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,7168,0.1249964502122667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,1024,0.013819555441538492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,10240,0.018333334061834548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,7168,0.021379555265108745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,512,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,128,0.01480977733929952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,6144,1.63463380601671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,32,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,2048,0.01482488877243466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,1024,0.011406222151385413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,3584,0.0343413319852617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,65536,0.07352444198396471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,768,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,3584,0.019578667150603402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,2560,0.021026665965716045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,3072,0.022582223018010456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,768,0.0314648879898919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,1536,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,3072,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,3584,0.008587555752860175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,512,0.005626666463083691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,8192,0.046888000435299344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,64,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,10240,0.020262221495310467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,768,0.0038808890514903595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,768,0.017197334104114108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,768,0.006355555521117316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,5120,0.02382755610677931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,256,0.016559110747443307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,10240,0.02558311157756382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,5120,0.11151911152733697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,4096,0.008339555727110969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,6144,0.04399289025200737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,32,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,3584,0.010062221851613786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,65536,0.1563431157006158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,768,0.017190222938855488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,768,0.007124444676770105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,10240,0.09226310915417141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,1536,0.006573333508438534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,768,0.014225777652528552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,2560,0.016914667354689706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,12288,0.03471111257870992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,4096,0.008019555773999956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,6144,0.017093333933088515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,65536,0.4349502192603217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,16384,0.04955288767814636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,1536,0.0092275556590822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,3072,0.08436977863311768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,65536,0.07798755831188627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,16384,0.0797591143184238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,10240,0.013800889253616333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,12288,0.024700444605615404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,512,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,4096,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,3584,0.020271999968422782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,65536,0.3234213458167182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,512,0.004228444563017951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,2048,0.01614755557643043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,128,0.010081777969996134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,1024,0.020958221620983545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,768,0.0037253333462609183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,1024,0.0151235560576121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,3072,0.031558222240871854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,7168,0.030606223477257624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,1024,0.004202666795916027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,32,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,64,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,16384,0.0280693338976966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,65536,0.24879378742641875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,1024,0.08903911378648545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,10240,0.0237128883600235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,7168,0.021528000632921856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,768,0.003974222060706881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,16384,0.13482844829559326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,3584,0.011222222612963783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,12288,0.019620445039537217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,5120,0.032243556446499295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,4096,0.012639111114872826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,2048,0.02558311157756382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,64,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,256,0.005924444645643234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,2560,0.0069022224181228215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,8192,0.46403201421101886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,6144,0.017472889688279893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,5120,0.023557333482636347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,12288,0.016921778519948322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,5120,0.23968709839714897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,3072,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,4096,0.029548443026012842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,4096,0.018553778529167175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,12288,0.03473688827620612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,10240,0.022978666755888198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,32,0.005499555418888728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,65536,0.11099111371570164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,4096,0.008258666429254744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,10240,0.035432888401879206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,16384,0.02514311174551646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,6144,0.050664888487921826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,6144,0.009336888790130615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,768,0.014541332920392355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,65536,0.5537297990587022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,5120,0.030474666092130873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,256,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,6144,0.009322666459613377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,1024,0.006206222292449739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,1024,0.004252444538805219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,4096,0.026365333133273657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2048,512,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,3072,0.016162667009565566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,3072,0.019017777509159513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,512,0.0052586665583981415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,7168,0.11074044307072957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,3072,0.007127111156781514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,2048,0.02708622150950962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,768,0.021690666675567627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,4096,0.01789066692193349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,1024,0.014488889111412896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,3584,0.019652444455358718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,8192,0.026389333936903212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,1536,0.00452977791428566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,128,0.005205333232879639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,768,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,1536,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,65536,0.4916666878594293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,65536,0.4708551300896539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,2048,0.008423999779754216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,3584,0.025219556358125474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,768,0.004238222208287981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,4096,0.015777778294351365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,768,0.017200888858901132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,8192,0.813336001502143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,7168,0.03749155667093065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,768,0.022475555539131165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,7168,0.059706668059031166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,768,0.014831110835075378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,5120,0.010078222387366826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,7168,0.033018665181265935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,1024,0.009306666751702627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,3584,0.010460444622569613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,128,0.019549333386951022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,7168,0.017491555876202054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,8192,0.022329777479171753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,2048,0.025156444973415796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,3072,0.011013333168294696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,32,0.003125333330697484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,10240,0.016709332664807636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,2560,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,256,0.00655377780397733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,1536,0.00889511075284746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,3072,0.018230222993426852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,10240,0.01682133310370975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,65536,2.8861707051595054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,8192,0.03779200050565932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,16384,0.6645084487067329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,768,0.021002666817771062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,32,0.007125332951545715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,16384,0.13087911076015896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,4096,0.2595546775394016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,3584,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,8192,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,32,0.0060159998635451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,32,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,8192,0.03366400135887994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,65536,0.21371911631690133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,128,0.020153777466879953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,4096,0.021881777379247878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3584,12288,0.38383110364278156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,6144,0.009702222214804756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,4096,0.013766222529941134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,5120,0.12197332912021214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,32,0.02922044528855218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,128,0.0028799999919202593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,64,0.006908444480763541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,2560,0.01663377715481652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,12288,0.014131555954615274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,64,0.005660444498062134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,3072,0.009645333720578088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,6144,0.02824444572130839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,16384,0.014985778265529208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,8192,0.012732444538010491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,3072,0.021947556071811255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,7168,0.09931377569834392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,1536,0.004561777744028303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,2048,0.016210666961140104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,768,0.004599110947714912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,16384,0.1330640051099989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,8192,0.033717334270477295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,10240,0.017883555756674874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,3072,0.02377066678471035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,65536,0.22578933503892687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,5120,0.0255386663807763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,65536,0.28466399510701496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,16384,0.05188888973659939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,10240,0.036559998989105225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,3584,0.04070933328734504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,7168,0.01686311099264357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,8192,0.013770666387346057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,1024,0.004488000025351842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,8192,0.020276443825827706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,1536,0.014819555812411837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,10240,0.01036000003417333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,10240,0.012349333200189801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,32,0.005939555664857228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,3584,0.029225776592890423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,16384,0.01976711054642995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,1536,0.01756177842617035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,16384,0.015484443969196744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,6144,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,16384,0.047028445535235934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,1024,0.013926222920417786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,256,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,128,0.009712888962692684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,1536,0.014807111687130399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,1024,0.008671999805503422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,10240,0.012758221891191272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,1536,0.01514933341079288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,1536,0.006647111227114995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,7168,0.014532445205582513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,16384,0.021010667085647583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,2560,0.008054222497675154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,4096,0.023354666100607976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,1536,0.010610666539933948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,4096,0.027436445156733196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,2560,0.010218666659461128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,32,0.002918222298224767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,65536,5.713331434461806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,256,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,5120,0.026757333013746474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,768,0.016172443826993305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,2048,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,768,0.009583999713261923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,3584,0.016142222616407607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,3584,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,3072,0.006629333313968446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,3584,0.2348088953230116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,3584,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,8192,0.01460444430510203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,4096,0.043756445248921715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,128,0.02401955591307746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,256,0.007983110845088959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,12288,0.01311822235584259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,4096,0.016872889465755887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,3072,0.013550221920013428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,1536,0.005277333160241445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,5120,0.01681777834892273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,512,0.01310488912794325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,3584,0.017438222964604694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,256,0.005665777872006099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,256,0.005631111148330901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,768,0.01681777834892273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,10240,0.028204444381925795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,1024,0.009874666730562845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,12288,0.02961155441072252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,16384,0.034712887472576566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,16384,0.045444445477591626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,16384,5.302630106608073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,6144,0.020632889535692003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,7168,0.05500444438722399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,256,0.009711110757456886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,1536,0.008126222425036961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,65536,0.1586808893415663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,10240,0.013985777894655863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,3584,0.01647111111217075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,4096,0.019648889700571697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,16384,0.03437422381507026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,512,0.015835555063353646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,3072,0.007626666790909237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,3072,0.016322667400042217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,512,0.004230222354332606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,16384,0.046125332514444985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,2560,0.010770666930410596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,10240,0.011767110890812345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,16384,0.06006133556365967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,2048,0.046279999944898814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,7168,0.04572711057133145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,1024,0.006959111326270633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,5120,0.018426666657129925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,2048,0.009003555609120263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,128,0.0052728888889153796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,4096,0.02073777715365092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,12288,0.013487111363146039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,8192,0.010665777656767102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,512,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,4096,0.045273777511384755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,768,0.03933155536651611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,128,0.007992000215583378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,512,0.012811555630630918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,512,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,512,0.01129066695769628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,1536,0.013461333182122974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,2560,0.019221333993805777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,4096,0.009701333112186855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,10240,0.044983112149768405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,5120,0.040138665172788836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,32,0.006627555522653792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,768,0.006040888941950268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,4096,0.006218666831652324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,7168,0.013879111243618859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,2560,0.02039377722475264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,10240,0.043567998541726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,8192,0.012075555821259817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,2560,0.024873778223991394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,2048,0.011492444409264458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,12288,0.05351822243796455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,5120,0.008310221963458592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,10240,0.1970826652314928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,1536,0.01074488874938753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,512,0.0031973332580592898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,512,0.021539555655585393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,16384,0.2906373341878255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,65536,3584,0.03647022114859687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,256,0.0035173334181308746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,10240,0.19781688849131265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,128,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,1536,0.006615110983451207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,1536,0.011166222393512726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,4096,0.022809776994917128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,8192,0.08453333377838135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,5120,0.00812533332241906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,12288,0.01791111131509145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,2048,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,1024,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,12288,0.08712444702784221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,3072,0.022053332792388067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,7168,0.0258568889564938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,12288,0.013507555756303998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,7168,0.04643733302752177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,1536,0.03423822257253859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,1536,0.008978666530715095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,2560,0.016029333074887592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,10240,0.007649777664078607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,3584,0.01074222226937612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,4096,0.008038221961922115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,3584,0.0913706686761644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,2560,0.011760888828171624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,6144,0.04442310995525784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,768,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,7168,0.027360000544124182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,1536,0.0070017774899800616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,65536,0.19544443819257948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,128,0.014470222923490735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,65536,0.08570044570498997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,1536,0.015816888875431485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,65536,0.06466577450434367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,768,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,3072,0.008716444174448649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,512,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,10240,0.03497599893146091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,3072,0.011376889215575324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,3072,0.10280888610415989
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,4096,0.041123555766211614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,512,0.014727999766667684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,7168,0.11949955092536078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,512,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,10240,0.011750222080283694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,256,0.015312888556056552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,2048,0.00608711110221015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,512,0.006114666660626729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,8192,0.012524444195959302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,256,0.005382222019963794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,4096,0.0170168893204795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,1536,0.010199111368921068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,768,0.003928889003064898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,768,0.005647111270162795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,3584,0.016540444559521146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,768,0.016062221593326993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,4096,0.01795111099878947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,12288,0.01720888912677765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,512,0.003548444559176763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,12288,0.050703111622068614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,1024,0.00976533359951443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,5120,0.00868444475862715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,4096,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,12288,0.019893333315849304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,6144,0.008947555389669206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,3584,0.02369244396686554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,3072,0.02462399999300639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,2048,0.008980444735950893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,7168,0.010053333308961656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,3072,0.013175110850069257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,3584,0.054457777076297335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,128,0.005884444548024072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,8192,0.016979555288950603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,3584,0.007313778003056844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,6144,0.009653333160612319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,3072,0.02492977844344245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,16384,0.02196088929971059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,5120,0.057414220439063184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,768,0.003563555578390757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,1536,0.006317333214812809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,1024,0.015456888410780164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,10240,0.3776888847351074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,65536,0.3067235681745741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,1024,0.004583999928500918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,512,0.0058444444504049085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,5120,0.009328888522254096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,2560,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,2048,0.009640889035330879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,128,0.015219555960761176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,3584,0.02853333287768894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,12288,0.024375110864639282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,1024,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,6144,0.033117334047953285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,5120,0.00868888861603207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,65536,0.05811466773351034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,512,0.00555644432703654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,1024,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,4096,0.017895110779338412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,2560,0.02793688906563653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,64,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,8192,0.02812533246146308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,2048,0.03980711102485657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,8192,0.030373334884643555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,12288,0.10001511043972439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,256,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,6144,0.02232622272438473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,4096,0.013705777625242868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,512,0.009597332941161262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,10240,0.04937511020236545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,1024,0.008166222108734978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,16384,0.027446221974160936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,10240,0.011735111474990845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,768,0.0035093331502543557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,512,0.0038275557259718576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,5120,0.006897777732875612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,12288,0.03768444392416213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,256,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,1024,0.013778666655222574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,128,0.00833422193924586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,5120,0.014153778553009033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,16384,0.03169066707293192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,768,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,4096,0.015136889285511441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,4096,0.018921777606010437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,1024,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,3072,0.01482399966981676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,5120,0.21067021952735054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,8192,0.02914666798379686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,16384,0.018636445204416912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,3072,0.010281778044170803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,2048,0.009339555270142024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,4096,0.018214222457673814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,2048,0.00871377769443724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,7168,0.009651555783218807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,64,0.04856444398562113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,1536,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,65536,3.2759306165907116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,768,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,768,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,16384,0.01719999975628323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,1024,0.00796088907453749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,1536,0.008323555191357931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,2560,0.009696000152164036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,7168,0.03584533267550998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,2560,0.008609777523411645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,65536,0.1847928894890679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,3072,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,16384,0.03694044550259908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,1024,0.014563555518786112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,7168,0.03152711192766825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,12288,0.017370666066805523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,6144,0.02269688910908169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,5120,0.01647822227742937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,128,0.006039111150635614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,2560,0.14348888397216797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,2560,0.04380177789264255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,10240,0.028561777538723413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,1536,0.006631111105283101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,12288,0.015471999843915304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,1024,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,10240,0.024076445235146418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,5120,0.012532444463835822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,4096,0.019848888119061787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,65536,0.06476444668240018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,10240,0.034436444441477455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,1024,0.006255110932721033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,6144,0.010661332971519895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,65536,0.05728799766964383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,6144,0.22019022040896943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,3584,0.00721777809990777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,5120,0.03142755561404758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,6144,0.03649866580963135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,16384,0.024146666129430134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,5120,0.03103111187616984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,256,0.03699200020896064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,1024,0.008957333034939235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,6144,0.020240889655219186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,16384,0.0385155545340644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,5120,0.053373333480623036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,6144,0.013089777694808112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,6144,0.02053955528471205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,10240,0.027552889453040227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,32,0.00309599998096625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,6144,0.04194844431347317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,3584,0.008002666963471307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,256,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,2048,0.010040889183680216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,5120,0.06545599963929918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,64,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,5120,0.010002666877375709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,4096,0.011366222467687396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,5120,0.008621333373917473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,3584,0.1334142287572225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,2560,0.017966222431924608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,7168,0.01961599952644772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,6144,0.01629333363638984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,8192,0.2802986568874783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,64,0.00572533326016532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,65536,0.16230844126807317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,4096,0.019482667247454327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,512,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,16384,0.01848088867134518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,16384,0.05042044321695963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,6144,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,5120,0.008685333033402761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,1024,0.008969777988062965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,3584,0.007612444460391998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,7168,0.01036888857682546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,5120,0.008709333009190029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,2560,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,7168,0.020558221472634208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,8192,0.015590222345458137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,16384,0.019094222121768527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,768,0.09502222140630086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,3584,0.008029333419269985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,3072,0.013806222213639153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,768,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,256,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,10240,0.06332621971766154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,10240,0.0903795560201009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,512,0.0069102222720781965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,16384,0.03040177623430888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,512,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,256,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,5120,0.006688888702127669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,6144,0.03605866763326857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,1024,0.01982222166326311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,128,0.006959111326270633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,4096,0.005555555638339784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,128,0.01110311100880305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,5120,0.014256889621416727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,7168,0.025145777397685583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,8192,0.015172445111804538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,768,0.004519999855094486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,64,0.007244444555706448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,4096,0.011143111520343356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,128,0.006303999986913469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,12288,0.13396267096201578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,1536,0.007642666498819987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,128,0.0038480001191298165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,7168,0.009376000199053023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,65536,0.06236355834537082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,1024,0.0052933332820733385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,256,0.01722311145729489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,5120,0.08228000005086263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,8192,0.011592888997660743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,8192,0.02743911080890232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,5120,0.006973333656787872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,10240,0.07008533345328437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,12288,0.1777413288752238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,12288,0.09368444151348537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,768,0.020000888241661918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,32,0.0058115555180443665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,3584,0.03069066670205858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,12288,0.04719822274314033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,1024,0.0042257776690853965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,32,0.005614222337802251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,768,0.01822933389080895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,128,0.004217777815130022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,128,0.002887111157178879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,8192,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,12288,0.031161778502994116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,1536,0.009659555223253038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,32,0.003555555724435382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,5120,0.016217778126398723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,8192,0.02071466710832384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,768,0.005626666463083691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,6144,0.006597333484225803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,6144,0.020980444219377305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,2560,0.010032888915803697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,64,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,768,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,2048,0.006604444649484422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,1536,0.02400355537732442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,1536,0.0225582222143809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,3584,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,8192,0.023987554841571387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,64,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,768,0.01755911111831665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,3584,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,32,0.006635555376609166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,1024,0.005973333285914526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,5120,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,6144,0.021198223034540813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,6144,0.015617777903874716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,4096,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,12288,0.014147554834683737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,1024,0.0042257776690853965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,3072,0.01649599936273363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,4096,0.04099911120202806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,5120,0.012076444096035428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,3072,0.037999110089408025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,3584,0.006228444476922353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,2048,0.018522666560279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,7168,0.017543110582563613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,768,0.017901332841979135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,65536,0.1400302251180013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,65536,6.475847032335069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,2048,0.02602400051222907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,32,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,128,0.011390222443474663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,7168,0.018932445181740653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,2048,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,2560,0.11457155810462104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,256,0.0052693333062860705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,8192,0.02311555544535319
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,4096,0.023567110300064087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,1536,0.020450666546821594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,6144,0.008964444200197855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,32,0.007670222057236566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,12288,0.01904088921017117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,5120,0.009761778016885122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,1024,0.008283555507659912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,3072,0.02383733292420705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,6144,0.022123555342356365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,6144,0.010667555862002902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,4096,0.010454222559928894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,2048,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,16384,0.017249777913093567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,5120,0.022058667408095464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,2048,0.01276088919904497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,512,0.003157333367400699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,1024,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,16384,0.02721600068940057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,3584,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,2560,0.009300444689061906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,10240,0.028900444507598877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,32,0.003568888952334722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,7168,0.012034667034943899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,1024,0.006529777828190062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,6144,0.14502844545576307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,16384,0.032223112053341336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,10240,0.011699555648697747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,256,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,65536,0.014103111293580798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,128,0.016388444436921012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,6144,0.00903555585278405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,12288,0.18861867321862114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,1536,0.024217777782016333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,512,0.003321777822242843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,16384,0.02881866693496704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,2048,0.012255999777052136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,1536,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,3072,0.019544000426928203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,6144,0.01037688884470198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,65536,0.0790897806485494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,5120,0.011917333636018964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,512,0.009654222263230218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,8192,0.010483555495738983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,3584,0.022230222821235657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,65536,0.06196710798475477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,128,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,12288,0.021867555048730638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,6144,0.012526222401195101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,10240,0.019475556082195707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,7168,0.024654222859276667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,10240,0.04966488811704847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,512,0.013990221752060784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,2560,0.010308444499969482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,10240,0.014685332775115967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,7168,0.012801777985360889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,4096,0.030576000610987347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,128,0.006210666563775804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,1536,0.020870221985710993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,4096,0.20461155308617485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,1536,0.004884444591071871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,10240,12.633637322319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,3072,0.012870222330093384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,6144,0.036277333895365395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,128,0.0041493334703975255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,2560,0.139338665538364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,768,0.007276444799370236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,3072,0.011335111326641507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,512,0.007668444679843054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,3584,0.016167110866970487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,512,0.0035297779573334586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,3072,0.007582222421964009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,10240,0.043255110581715904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,3584,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,512,0.012268444730175866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,3072,0.02840533190303379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,3072,0.016809778081046212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,512,0.011768889096048145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,4096,0.00847822262181176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,4096,0.08967644638485378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,512,0.01685688893000285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,128,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,10240,0.08997599946127997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,2048,0.014055111342006259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,128,0.015185778339703878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,8192,0.011361777782440186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,6144,0.026150221625963848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,7168,0.08532444636027019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,3072,0.010112889111042023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,10240,0.016216889023780823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,8192,1.7517147064208984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,256,0.005191111316283544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,256,0.009912889036867354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,5120,0.3910337818993463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,2560,0.016175111134847004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,2048,0.020862221717834473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,2560,0.010072000324726105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,1024,0.0058853332367208265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,2560,0.010328888893127441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,12288,0.02407111061943902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,1536,0.004516444272465176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,65536,0.14737244447072348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,12288,0.0407262212700314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,1024,0.02304977840847439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,256,0.007301333049933116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,12288,0.48152531517876523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,6144,0.04918933245870802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,1024,0.00907733374171787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,512,0.07712444331910875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,32,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,10240,0.027101332942644756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,3584,0.007556444240940942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,512,0.012969777815871768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,3584,0.014465777410401238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,4096,0.04028444488843282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,4096,0.01811288959450192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,12288,0.06403022342258029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,65536,0.07447999715805054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,10240,0.021007999777793884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,16384,0.021730666359265644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,1536,0.013736888766288757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,5120,0.008360889222886827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,32,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,10240,0.01235466698805491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,2560,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3584,16384,0.036496887604395546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,1536,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,5120,0.04938933253288269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,1536,0.004528000122971005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,4096,0.01070666644308302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,32,0.0029048888633648553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,3072,0.08399822314580281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,10240,0.017098666893111337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,6144,0.026111111044883728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,3584,0.02868888775507609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,3584,0.01995466649532318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,512,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,128,0.003855111284388436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,1536,0.006619555668698416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,512,0.003896000070704354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,512,0.006959111326270633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,5120,0.015344889627562629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,8192,0.010683555569913654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,16384,0.05598755677541097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,1536,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,2560,0.02419022222359975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,128,0.005902222047249476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,2560,0.005872000008821487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,6144,0.010141332944234213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,1024,0.038022223446104265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,4096,0.011848889291286469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,4096,0.027190221680535212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,8192,0.014914666612943014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,3584,0.008356444537639618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,4096,0.0103493332862854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,10240,0.03530399998029073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,10240,0.026877333720525105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,1536,0.0147706667582194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,32,0.06316800249947442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,2048,0.010071111222108206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,6144,0.06346666812896729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,16384,0.019944889677895438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,768,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,64,0.06193955739339193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,5120,0.015644444359673392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,4096,0.011365333365069495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,3072,0.007736889024575551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,7168,0.2795635594262017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,3584,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,1536,0.022294221652878657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,4096,0.014526221487257214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,768,0.018231110440360177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,12288,0.039545777771208025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,16384,0.03131555517514547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,5120,0.033944000800450645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,7168,0.018975999620225694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,1024,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,3584,0.008351999852392409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,6144,0.03919911053445604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,2048,0.011004444625642566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,5120,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,6144,0.010378667049937779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,65536,0.05850311120351156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,4096,0.015108444624476962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,32,0.0028124445428450904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,2560,0.011718221836619906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,7168,0.034400890270868935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,6144,0.025378665990299646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,2048,0.012453333371215396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,6144,0.035460445615980364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,256,0.00518222235971027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,3584,0.02576711111598545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,3072,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,1536,0.004604444321658877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,65536,0.08223288589053683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,768,0.015307555596033732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,7168,0.02164622147878011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,512,0.016145777371194627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,512,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,768,0.016932444439993966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,16384,4.498182084825304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,256,0.010332444475756751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,768,0.007668444679843054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,512,0.019627556204795837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,32,0.006979555719428592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,4096,0.010068444742096795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,768,0.018812444474962022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,6144,0.033655110332700945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,8192,0.016391999191708036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,4096,0.02313599983851115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,12288,0.018357333209779527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,2560,0.00702044450574451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,8192,0.016287111573749118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,10240,0.07688622342215644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,256,0.004923555586073133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,6144,0.011266666981909009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,1024,0.0038951109680864546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,12288,0.06921777460310194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,65536,0.3896622127956814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,1024,0.006933333145247565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,256,0.015167999598715039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1024,64,0.0059164443777667145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,8192,0.022319111559126113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,64,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,16384,0.027783112393485174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,6144,0.09255644347932603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,32,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,5120,0.11074311203426784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,2048,0.007324444750944774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,8192,0.01538755496342977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,128,0.01038577738735411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,512,0.04237066706021627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,3072,0.06613777743445502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,8192,0.051148444414138794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,128,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,64,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,10240,0.024296000599861145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,8192,0.07162400086720784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,64,0.0035422220826148987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,5120,0.0121742222044203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,3072,0.6268346574571397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,10240,0.02458577851454417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,256,0.009381333159075843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,1536,0.006688888702127669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,1024,0.008972444468074376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,768,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,128,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,8192,0.012356444365448423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,6144,0.009400000174840292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,5120,0.012834666503800286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,4096,0.007339555356237624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,12288,0.023642667465739783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,8192,0.022282666630215112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,3072,0.007278222176763747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,3072,0.01001422190003925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,1024,0.011693333586057028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,4096,0.010206222534179688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,6144,10240,0.032893333170149065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,128,0.004874666531880696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,6144,0.009644444617960189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,12288,0.0159262220064799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,3584,0.009354666703277165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,5120,0.019505777292781405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,768,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,10240,0.04412710997793409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,65536,0.08131021923489042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,3584,0.015178667174445258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,12288,0.01316177762216992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,8192,0.010673777924643623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,768,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,1536,0.006234666539563074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,8192,0.042640888028674655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,3072,0.06117510795593262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,3072,0.007891555627187094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,12288,0.037371555964152016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,1536,0.004918222212129169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,768,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,512,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,512,0.0032426667296224167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,512,0.01461155547036065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,8192,0.38129422399732804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,2048,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,3072,0.02457333273357815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,128,0.016144000821643405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,128,0.011721777419249216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,65536,0.0663599967956543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,8192,0.028034667174021404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,1536,0.01071733319097095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,64,0.0029235554652081597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,4096,0.023595554961098567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,1024,0.007307555940416124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,1536,0.004593777573770947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,10240,0.029959112405776978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,65536,0.07551466756396823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,10240,0.025397333833906386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,8192,0.12762578328450522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,65536,0.19985422823164198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,16384,0.055481778250800244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,2560,0.029692444536421034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,10240,0.009977777798970541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,2048,0.00630044440428416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,7168,0.01073155552148819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,3584,0.058113780286577016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,16384,0.011583999627166323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,65536,0.1497404442893134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,6144,0.018629334039158292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,256,0.006217777729034424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,5120,0.01148622234662374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,512,0.005571555760171678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,12288,0.19968799750010172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,1536,0.005724444571468566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,3072,0.03490044342146979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,256,0.014825777875052558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,4096,0.011787555283970304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,3072,0.03876533442073398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,2048,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,2560,0.026333333717452154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,1536,0.012248000337017907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,16384,0.01941155559486813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,512,0.005241777747869492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,512,0.013743999931547375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,2048,0.02777155571513706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,5120,0.04003289010789659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,256,0.009358222285906473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,8192,0.1197626723183526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,64,0.008004444340864817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,2048,0.020813332663642038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,4096,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,6144,0.02128533356719547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,3072,0.0069164443347189165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,1024,0.005558222118351195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,128,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,5120,0.02723199956946903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,768,0.00565244464410676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,4096,0.008009778128729926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,7168,0.019903111788961623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,2560,0.01587022178702884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,12288,0.016882666283183627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,10240,0.012759110993809171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,6144,0.014148443937301636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,2048,0.017214222086800467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,1536,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,128,0.011736888852384357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,1536,0.0162124451663759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,768,0.020051555501090158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,1536,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,16384,0.019677332705921598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,1536,0.015488000379668342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,6144,0.010155555274751451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,16384,0.052413334449132286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,7168,0.03846222162246704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,65536,0.21528889073265922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,2560,0.006867555694447623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,128,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,16384,0.029919111066394385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,16384,0.007972444097201029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,2048,0.014753777119848462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,4096,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,5120,0.020455110404226515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,768,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,1536,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,5120,0.05041244294908312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,12288,0.10153777731789483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,7168,0.02365422248840332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,4096,0.008667555948098501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,4096,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,65536,0.1093662248717414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,8192,0.019298666053348117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,10240,0.016382222374280293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,7168,0.03053155541419983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,128,0.014440000057220459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,2560,0.023012444376945496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,1024,0.010793777803579966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,5120,0.011012444065676795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,768,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,768,0.022410665949185688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,64,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,3072,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,1024,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,1536,0.011056889262464313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,2048,0.04490577843454149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,65536,0.1391635603374905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,16384,0.01718311177359687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,1024,0.0038728887836138406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,3584,0.018239110708236694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,512,0.0035217776894569397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,1536,0.02230666743384467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,512,0.013610666824711693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,65536,0.06227022409439087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,4096,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,7168,0.06266044245825873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,512,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,3072,0.012424888710180918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,2048,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,8192,0.023911999331580266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,32,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,7168,0.07472444242901273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,2560,0.015862221519152325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,4096,0.01721155477894677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,256,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,3584,0.008319999608728621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,5120,0.15864533848232693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,1024,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,16384,0.02239022155602773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,2560,0.015091554986106025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,12288,0.04530133472548591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,12288,0.022270222504933674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,6144,0.028780443800820246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,5120,0.009317333499590555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,256,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,3584,0.010686222049925061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,5120,0.055567999680837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,2560,0.030577778816223145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,2560,0.013471999930010902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,3072,0.011870221959220039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,512,0.0166666673289405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,256,0.005589333259397083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,1536,0.010752889017264048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,2560,0.0151591118839052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,16384,0.02849510974354214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,4096,0.00868444475862715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,12288,0.061977777216169566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,1536,0.024409777588314478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,256,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,32,0.0034391110142072043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,16384,0.12596710522969565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,128,0.01647644407219357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,3584,0.02845422095722622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,1024,0.009071110851234859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,12288,0.00996088898844189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,32,0.006980444822046492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,10240,0.008998221821255153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,3072,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,512,0.01978577839003669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,128,0.014496889379289416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,768,0.0037120001183615792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,768,0.003564444267087512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,128,0.013079110946920184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,768,0.007428444094128079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,5120,0.06382489204406738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,768,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,1024,0.003899555653333664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,3072,0.08427288797166611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,256,0.004587555511130227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,3072,0.15243644184536403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,1536,0.036664889918433294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,3584,0.021002666817771062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,12288,0.030577778816223145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,32,0.006566222343179915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,2048,0.021151999632517498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,3072,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,6144,0.008942222429646386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,128,0.014509333504570855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,2048,0.03629422187805176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,65536,0.1258044507768419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,32,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,65536,0.0998106665081448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,65536,0.057322667704688184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,32,0.007213333414660559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,2048,0.006917333437336816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,3584,0.019713777634832595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,5120,0.008339555727110969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,32,0.003044444446762403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,2048,0.006829333388143116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,4096,0.014903111590279473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,7168,0.07422044542100695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,5120,0.01519466605451372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,1536,0.00665866666369968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,3072,0.01382844481203291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,16384,0.04182844360669454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,768,0.0039048890272776284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,768,0.007696888513035244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,256,0.01722577710946401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,256,0.031075557072957356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,6144,0.01330044451687071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,7168,0.013589333328935834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,65536,0.18886133035024008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,3584,0.008344000412358178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,2048,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3072,256,0.00565422202150027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,6144,0.0118568887313207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,8192,0.013912889692518445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,1024,0.014158222410413953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,3584,0.055211553970972695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,1536,0.00453599997692638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,64,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,3072,0.013886222408877479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,12288,0.01721955504682329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,3072,0.015837333268589444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,3072,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,2048,0.010472888747851053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,12288,0.008151111503442129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,8192,0.0200311111079322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,7168,0.028491556644439697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,4096,2048,0.026213333010673523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,512,0.013429332938459186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,512,0.006033777776691649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,65536,2.6537999047173395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,12288,0.028148445818159316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,1536,0.007343110938866933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,65536,0.02535733415020837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,2560,0.018887110882335238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,65536,9.84945593939887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,3072,0.007599111232492659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,3584,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,2560,0.016149333781666227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,256,0.014484445254007975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,128,0.03696799940533108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,7168,1.9033618503146703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,1024,0.023733332753181458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,2048,0.029478222131729126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,10240,0.010059555371602377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,8192,0.008307555483447181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,65536,0.9001519944932727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,64,0.0033235556135574975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,10240,0.015107555521859063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,32,0.003535110917356279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,256,0.02095555596881443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,1536,0.008303110798199972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,6144,0.26045065455966526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,4096,0.010323555933104621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,2560,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,256,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,6144,0.018556444181336295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,32,0.013759111364682516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,8192,0.02150933278931512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,3072,0.0090506664580769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,1024,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,3584,0.013777777552604675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,4096,0.009226666556464301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,32,0.005394666559166378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,5120,0.01761511133776771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,32,0.005612444546487596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,2048,0.010056888891590966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,16384,0.03964089022742377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,768,0.009382222261693742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,65536,0.02370044423474206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,5120,0.00831911133395301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,128,65536,0.07773066891564263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,2560,0.020036444067955017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,1024,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,16384,0.04012711180580987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,128,0.004906666775544484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,6144,0.020580444071027968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,16384,0.08309688833024767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,3584,0.026183999247021143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,16384,0.3528302245669895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,32,0.0024871110088295406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,12288,0.852416886223687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,4096,0.010890666809346942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,3072,0.024899555577172175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,768,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,4096,0.02367288867632548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,768,0.004898666507667965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,2048,0.15443021721310085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,256,0.005977777971161737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,4096,0.010990222295125326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,768,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,1536,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,3584,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,12288,0.06388355626000299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,12288,0.03682755430539449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,3584,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,256,0.016373333003785875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,2560,0.012651555240154266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,1536,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,512,0.014811555544535318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,8192,0.18130666679806182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,128,0.003126222226354811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,128,0.014782221780882942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,7168,0.018237334158685472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,2560,0.013464888764752282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,16384,0.10808889071146648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,10240,0.012368888490729861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,1536,0.034620443979899086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,1536,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,768,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,256,0.006903999795516332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,7168,0.011355555719799466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,65536,0.016499555773205228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,3072,0.015451555450757345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,65536,0.0489671097861396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,3584,0.026797332697444495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,3584,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,3584,0.023200000325838726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,32,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,5120,0.030749334229363337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,768,0.003624888757864634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,3072,0.0814888874689738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,1024,0.0058337777025169795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,2560,0.010517333116796283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,768,0.02073777715365092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,12288,0.020834666159417894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,1536,0.01960088809331258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,768,0.005858666780922148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,2560,0.007055111229419708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,1536,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,64,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,6144,0.009646221995353699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,1024,0.005244444641802046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,768,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,256,0.01791111131509145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,3584,0.4029688835144043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,8192,0.010715555813577441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,512,0.005998222364319696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,12288,0.01460444430510203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,3072,0.04272622201177809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,64,0.0331857767370012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,4096,0.007582222421964009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,3072,0.08297955327563815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,10240,0.012160888976520963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,16384,0.01070844464831882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,32,0.005963555640644497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,16384,0.01981244484583537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,5120,0.01962311069170634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,65536,0.14634488688574895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,2560,0.00703288863102595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,2560,0.015127110812399121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,128,0.022107554806603327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,7168,0.014820444915029736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,1024,0.01687111126052009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,8192,0.04611022273699442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,64,0.0038942222793896994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,5120,0.024455999334653217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,7168,0.014323555760913424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,256,0.011033777561452655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,16384,0.2289724349975586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,2560,0.02514044443766276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,1536,0.006607111129495833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,128,0.0028897778441508612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,768,0.03423110975159539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,65536,0.4929742283291287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,12288,0.01816444430086348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,10240,0.15371199448903403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,32,0.0029155556112527847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,2560,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,2560,0.02831733226776123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,6144,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,2560,0.006614222294754452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,1024,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,3584,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,256,0.00565422202150027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,32,0.0028035555862718155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,6144,0.1279075543085734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,65536,0.07527110973993938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,512,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,2048,0.0642755561404758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,512,0.01314399970902337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,128,0.011783999701340994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,2048,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,128,0.01277599980433782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,2048,0.005192888693677055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,5120,0.02431022293037838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,256,0.015300444430775113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,256,0.005375999957323074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,768,0.004308444344335132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,512,0.00526577772365676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,3584,0.014965333872371249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,12288,0.023542222049501207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,2048,0.01406222250726488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,1024,0.004278222305907143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,4096,0.0985671083132426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,768,0.013165333204799227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,3072,0.0075431110130416016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,6144,0.018253333038753934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,128,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,3584,0.03871377640300327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,12288,0.05007822314898173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,256,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,2560,0.030369778474171955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,768,0.004607111215591431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,10240,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,1024,0.014404444230927361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,2048,0.007236444287829929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,2048,0.006311999840868845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,768,0.005958222266700532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,768,0.016136000553766888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,10240,0.30495643615722656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,256,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,128,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,6144,0.023499555057949487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,12288,0.03110044532352024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,128,0.015054222610261707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,3584,0.029367112451129492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,1024,0.013807111316257052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,1024,0.005919999960396025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,7168,0.02881066666709052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,128,0.011266666981909009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,256,0.0052382221652401825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,768,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,3584,0.0069155556460221606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,5120,0.04836977852715386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,4096,0.009319999979601966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,12288,0.04020266731580099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,3584,0.009705777797434065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,768,0.009343999955389235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,65536,0.062441779507531055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,65536,0.8936568366156684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,65536,0.3244640032450358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,65536,0.10152000188827515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,12288,0.02096533278624217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,32,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,64,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,2560,0.01219733307758967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,1024,0.005899555567238066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,2560,0.015136889285511441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,1024,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,3072,0.016914667354689706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,7168,0.009696000152164036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,512,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,3072,0.018762666318151686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,32,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,7168,0.010863111250930362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,4096,0.015942222542232938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,7168,0.09407733546362983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,10240,0.04254755708906385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,1536,0.11575999524858262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,2560,0.017678222722477384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,64,0.005615111026499007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,1536,0.014463999205165438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,4096,0.010185778141021729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,6144,0.04595999916394552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,64,0.0035111109415690103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,12288,0.046167999505996704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,32,0.011408000356621213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,7168,0.010030222435792288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,3584,0.014190221826235453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,12288,0.021708443760871887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,7168,0.020237333244747586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,8192,0.015178667174445258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,128,0.0031146665828095544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,1536,0.018943111101786297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,128,0.021755556265513103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,1024,0.005730666634109285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,10240,0.032460444503360324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,2048,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,5120,0.011983111500740051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,32,0.010077333284748925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,5120,0.016571554872724745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,10240,0.025747555825445387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,2560,0.009829333258999718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,8192,0.023418666587935552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,8192,0.020215999748971727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,128,0.0052800000541739995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,256,0.1772915522257487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,3072,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,2048,0.008791999684439765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,1024,0.019346666004922655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,768,0.015357333752844067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,1024,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,16384,0.04123911261558533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,4096,0.009009777671760982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,5120,0.36953245268927676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,3072,0.03683200147416856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,32,0.007632888853549957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,16384,0.024101333485709295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,6144,0.030889779329299927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,7168,0.0196088883611891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,65536,0.07033689154518975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,5120,0.010088889135254754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,2560,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,1536,0.019316444794336956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,2560,0.01583377851380242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,128,0.011760888828171624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,3072,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,12288,0.045385777950286865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,8192,0.03351555599106683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,6144,0.00998488896422916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,2560,0.04991733365588718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,7168,0.010052444206343757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,1024,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,256,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,128,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,2048,0.0052133335007561584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,64,0.005544888890451855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,64,0.0026026666164398193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,3072,0.19178311030069986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,10240,0.017264000243610807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,3072,0.015864888827006023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,128,0.006309333360857434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,16384,0.07337244351704915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,6144,0.009210666848553551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,1536,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,6144,0.01740533279048072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,256,0.014801777071423002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,256,0.01479200025399526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,8192,0.01035822182893753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,4096,0.014365333649847241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,64,0.0026400000270869997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,3072,0.017749332719379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,768,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,8192,0.011052444577217102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,8192,0.05784444676505195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,32,0.009456000394291347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,1536,0.017869333426157635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,5120,0.24707643191019693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,16384,0.6063689125908746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,8192,0.04736622174580892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,4096,0.07451555464002821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,12288,0.04981777734226651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,3072,0.013539555172125498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,1536,0.02236622240808275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,7168,0.0162471118900511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,65536,0.5301991038852268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,768,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,3584,0.00902133352226681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,2560,0.029778665966457788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,256,0.0064604443808396654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,6144,0.018908444378111098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,10240,0.019903999235894945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,7168,0.03235377868016561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,128,0.004933333231343163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,4096,0.012147555748621622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,6144,0.013207111093733044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,5120,0.008635555704434713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,1536,0.006299555715587404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,6144,0.04469866553942362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,12288,0.03572355707486471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,65536,0.062450667222340904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,128,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,10240,0.01516355574131012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,1024,0.0058577776783042485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,4096,0.01684177749686771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,12288,0.02071288890308804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,4096,0.014910222755538093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,3584,0.00795288880666097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,16384,0.02682311170630985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,7168,0.1078631083170573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,1024,0.009069333473841349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,1024,0.004204444587230682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,3584,0.04915822214550442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,2048,0.018185777796639335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,6144,0.031719999180899724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,128,0.0037377778854635027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,128,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,1024,0.004540444248252445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,65536,1.9067314995659723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,3584,0.09825688600540161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,3072,0.007000889215204451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,256,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,12288,1536,0.0299911101659139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,4096,0.009683555199040307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,64,0.021026665965716045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,6144,0.02125866711139679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,512,0.02160888910293579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,128,0.012792888614866467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,8192,0.018760888112915885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,2560,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,768,0.06329244375228882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,12288,0.013168888787428537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,8192,0.24010843700832793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,16384,0.03484266665246751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,4096,0.12319821781582302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,1536,0.045422222879197865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,10240,0.02373955481582218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,12288,0.01792888840039571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,16384,0.03599288728502061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,10240,0.012311999996503195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,7168,0.30997244517008465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,1024,0.0061928890645504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,1536,0.006746666712893381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,5120,0.01293333371480306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,8192,0.007984889050324758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,7168,0.023183110687467787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,6144,0.010071111222108206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,768,0.007607999775144789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,768,0.004519999855094486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,512,0.012442666623327466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,5120,0.02091111077202691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,6144,0.01959466603067186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,1536,0.004564444637960858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,6144,0.07883555359310575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,65536,0.06726844443215264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,65536,0.08322399854660034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,128,0.006225777582989798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,16384,0.022667555345429316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,7168,0.03713511096106635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,4096,0.023037332627508376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,256,0.0069288888739215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,6144,0.014099554883109199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,65536,0.060903999540540904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,3584,0.015527110960748462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,5120,0.01847644481394026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,3584,0.020412445068359375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,3584,0.008301333420806462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,7168,0.025638222694396973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,1024,0.013164444102181328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,4096,0.01070577816830741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,12288,6144,0.04702933298216926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,16384,0.0557804438802931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,6144,0.00657066661450598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,128,0.008447111480765873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,256,0.017119111286269296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,2048,0.006691555596060223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,2048,0.009663999908500248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,4096,0.013524444566832649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,1024,0.017255999975734286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,5120,0.014114666316244336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,3584,0.023668444818920557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,256,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,2048,0.017539555827776592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,768,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,768,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,12288,0.03711555401484171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,2048,0.010038221875826517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,1536,0.011133333047231039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,10240,0.011400889191362592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,3072,0.01242844429281023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,512,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,3584,0.05837955739763048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,65536,0.30993689431084526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,12288,0.02441066669093238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,256,0.005183111048407025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,6144,0.011367999845080905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,1024,0.01921688848071628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,12288,0.028040889236662123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,2560,0.013813333378897773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,2048,0.013772444592581855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,3072,0.010377777947319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,256,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,6144,0.013075555364290873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,65536,0.07961510949664645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,5120,0.010774222513039907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,4096,0.007989333735571967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,3584,0.1067662239074707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,2048,0.01609599921438429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,64,0.0058497778243488735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,3072,0.005528888768619961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,10240,0.04030311107635498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,65536,0.05770488580067953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,4096,0.017139555679427255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,2048,0.01036444471942054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,128,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,5120,0.00960888879166709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,10240,0.014612444572978549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,5120,0.008967999782827165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,2048,0.010602666272057427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,6144,0.009857777920034196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,1024,0.0058764442801475525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,512,0.009666666388511658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,2048,0.028625776370366413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,7168,0.010235555469989777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,512,0.10080622302161323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,5120,0.016954667038387723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,10240,0.02642311155796051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,2560,0.009171555439631144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,4096,0.019409777389632333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,768,0.016788444585270353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,5120,0.011144888897736868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,8192,0.01035377797153261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,512,6144,0.007353777686754863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,3072,0.02667288813326094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,8192,0.03204177816708883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,5120,0.017196445001496207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,16384,0.040707555082109236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,256,0.0029084444459941653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,10240,0.08323999908235338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,512,0.004271999829345279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,10240,0.01588088936275906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,3072,0.04827377862400479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,10240,0.0749866697523329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,5120,0.015501333607567681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,2048,0.033260444800059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,3072,0.016120889120631747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,768,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,7168,0.029508444997999404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,2048,0.00638133328821924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,32,0.0028906667398081887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,12288,0.21796533796522352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,256,0.0035013332962989807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,65536,0.19112622737884521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,7168,0.02365511159102122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,8192,0.048483557171291776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,1536,0.014507555299335055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,1536,0.007415999968846639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,2048,0.014459555347760519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,32,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,1536,0.004546666724814309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,1536,0.006593777901596493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,2560,0.013134222063753339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,64,0.005557333429654439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,7168,0.02665422194533878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,7168,0.13370400004916722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,2560,0.03197244471973843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,2048,0.0069724445541699724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,10240,0.0144533332851198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,256,0.015160888433456421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,5120,0.014338667194048563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,8192,0.01722133325205909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,32,0.005551110953092575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,12288,0.013432888521088494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,128,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,3072,0.012519111235936483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,1536,0.0076248885856734375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,5120,0.033570667107899986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,256,0.0069191112286514705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,256,0.02685955497953627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,16384,0.019154666198624503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,256,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,128,0.005247999810510212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,4096,0.026543110609054565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,128,0.005388444496525659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1536,65536,0.03732622332043118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,8192,0.014931556251313953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,7168,0.009655999640623728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,2560,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,512,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,16384,0.01514311134815216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,5120,0.008969777988062965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,3584,0.007689778175618913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,768,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,4096,0.007996444072988298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,512,0.011039111349317761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,10240,0.017216889394654166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,3072,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,512,0.018339556124475267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,2048,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,256,0.0061048890153567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,7168,0.01534844438234965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,6144,128,0.00349866681628757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,7168,0.25071999761793345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,10240,0.15391822655995688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,10240,0.01167022188504537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,5120,0.011384888655609555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,3072,0.0069822221994400024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,7168,0.014326221413082547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,512,0.0063457778758472866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,2048,0.00552977787123786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,3072,0.009081777599122789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,5120,0.015403555499182807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,7168,0.03532977898915609
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,6144,0.00868711123863856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,65536,0.06258666515350342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,64,0.003173333282272021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,7168,0.025103110406133864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,1024,0.018225777480337355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,128,0.005606222069925732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,7168,0.014239999983045789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,128,0.1998044384850396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,6144,0.20077778233422172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,16384,0.15085777971479628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,1536,0.005444444302055571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1536,10240,0.011907555162906647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,1536,0.01978222197956509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,7168,0.013276444541083442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,10240,0.04325777623388502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,8192,0.02405333353413476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,32,0.005574222240183089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,2560,0.011168888873524137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,768,0.0041644444896115195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,3584,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,128,0.005872000008821487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,16384,0.02230577833122677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,6144,0.04233333468437195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,12288,13.90413072374132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,1024,0.010695111420419482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,8192,0.036312000619040594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,12288,0.01475644442770216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,5120,0.029510223203235205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,3072,0.01073333372672399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,10240,0.04236800140804715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,1536,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,1536,0.45780356725056964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,32,0.003185777821474605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,2560,0.009320889082219865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,5120,0.02816533380084568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,4096,0.025046222739749487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,5120,512,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,1024,0.01924000018172794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,768,0.009424000150627559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,12288,0.03046400017208523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,7168,0.013359111216333179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,1024,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,12288,0.021359110871950786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,6144,0.04624799887339274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,2560,0.012379555238617791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,512,0.04872622092564901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,6144,64,0.005532444351249271
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,16384,0.02088533341884613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,1024,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,2560,0.013092444174819522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,2048,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,128,0.009665778113736046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,1536,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,768,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,16384,0.028407110108269587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,128,0.014516444669829475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,1536,0.01388533330625958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,4096,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,256,0.009511111511124505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,1536,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,16384,0.02072533302836948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,6144,0.009395555489593083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,64,0.00281866660548581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,3072,0.021838222940762837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,512,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,64,0.003156444471743372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,6144,0.009353777600659264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,1536,0.004864000197913912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,2048,0.005232000102599462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,65536,0.09439733293321396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,128,0.003129777808984121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,3584,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,3072,0.014535999960369535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,6144,0.017094223035706412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,768,0.0069164443347189165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,4096,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,65536,0.025469332933425903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,1024,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,2048,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,6144,0.009362666971153682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,2048,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,6144,0.01314399970902337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,6144,0.0251511103577084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,16384,0.03781066669358148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,128,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,2560,0.09113333622614543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,3072,0.019256000717480976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,6144,0.07822399669223361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,128,0.002891555635465516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,128,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1536,2560,0.01405244403415256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,128,0.005203555441564984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,3072,0.015856888559129503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,4096,0.09647911124759251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,128,0.03296444482273526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,3584,0.01757244434621599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,128,0.006642666541867786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,1024,0.004541333350870344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,2560,0.007159111400445302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,12288,0.014840889308187695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,7168,0.015772443678643968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,12288,0.011387555963463254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,8192,0.01033688916100396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,16384,0.018963555494944256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,1536,0.015519110692871941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,768,0.014127110441525778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,65536,0.2349697748819987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,3584,0.019679110911157396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,3072,0.06510666343900892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,6144,0.026657778355810378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,256,0.005673777725961473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,16384,0.6330782042609321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,2560,0.013796444568369122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,3072,0.011537777880827585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,6144,0.023810666468408372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,5120,0.06997422377268474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,5120,0.02869510981771681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,8192,0.03901244534386529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,768,0.01722311145729489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,1536,0.013566222455766467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,512,0.0040942223535643685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,128,0.0053022222386466126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,3584,0.008440000315507254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,2048,0.011583111352390714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,768,0.005968889014588461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,65536,0.06142933501137627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,5120,0.030504888958401147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,1024,0.005251555393139522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,32,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,512,0.006917333437336816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,1536,0.024865777956114873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,64,0.0038924444880750445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,12288,0.027549333042568628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,256,0.015196444259749519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,5120,0.011767110890812345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,12288,0.02441066669093238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,1536,0.006248888870080312
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,6144,0.12978222635057238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,1024,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,1536,0.006559999866618051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,128,0.006204444501135085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,1536,0.004940444396601783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,2560,0.018145778112941317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,3584,0.00997866690158844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,768,0.015912888778580558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,1024,0.009478222164842818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,65536,0.06566044357087877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,512,0.006576889091067844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,512,0.01310222182008955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,12288,0.022156443860795762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,1024,0.005254222287072076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,64,0.0032231110251612137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,3072,0.04117066661516825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,2048,0.006754666566848755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,2560,0.025964443882306416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,12288,0.027836445305082533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,7168,0.01854311095343696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,3072,0.007655999726719326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,256,0.027832001447677612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,5120,0.011775111158688864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,12288,0.02536800007025401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,12288,0.01881511178281572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,2048,0.022671111755900916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,8192,0.13864178127712673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,768,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,128,0.09522133403354222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,256,0.016531555189026725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,8192,0.02134311033619775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,2560,0.7487617598639594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,2560,0.007612444460391998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,5120,0.017664889494578045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,1024,0.007648000286685095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,64,0.00962400022480223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,256,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,4096,0.007991111112965478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,10240,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,768,0.0158595558669832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,2048,0.006616000086069107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,1536,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,128,0.010068444742096795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,65536,4.747865041097005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,4096,0.02288444505797492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,16384,0.027991112735536363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,4096,0.015517334143320719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,10240,0.0413697792424096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,7168,0.08815555440054999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,3072,0.03744622071584066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,10240,0.022622222701708477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,7168,0.4611031214396159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,3072,0.0513591104083591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,12288,0.08309866984685262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,3584,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,128,0.005941333456171884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,10240,3.090936024983724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,12288,0.13817066616482204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,5120,0.04495111107826233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,2560,0.014673777752452426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,5120,16384,0.042506668302747935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,1024,0.011355555719799466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,5120,0.016488000750541687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,256,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,2560,0.0514613323741489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,1536,0.018567999203999836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,1536,0.03445600138770209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,6144,0.009638222555319468
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,2048,0.02738933265209198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,768,0.006490666833188798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,1024,0.011749332977665795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,16384,128,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,256,0.010760000182522668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,4096,0.06698666678534614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,128,0.003961777935425441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,2560,0.04826044374042087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,16384,0.046097778611712985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,1024,0.015185778339703878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,512,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,7168,0.07502310805850558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,16384,0.032003555032942034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,3584,0.015817777978049386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,2560,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,65536,2.4078826904296875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,7168,0.08401511112848918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,2048,0.01788444485929277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,8192,0.02732177740997738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,3584,0.014817777607176038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,12288,0.02527644402450985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,12288,0.013462222284740873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,2048,0.009397333694828881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,1536,0.014127999544143677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,4096,0.01722577710946401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,4096,0.11480532752143012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,16384,0.019751111666361492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,7168,0.019626667102177937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,4096,0.00870488915178511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,5120,0.030054221550623577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,8192,0.010718222293588849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,10240,0.028756442997190688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,128,0.010035555395815108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,512,0.017204445269372728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,8192,0.06534133354822795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,3072,0.008744888835483128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,768,0.016191111670600045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,5120,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,2560,0.020957332518365648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,8192,0.025111999776628282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,8192,0.01719199948840671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,512,0.012936000194814471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,3584,0.0069208890199661255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,3584,0.03779022230042352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,2560,0.014140443669425117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,2560,0.00978133330742518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,6144,0.0649155577023824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,3072,0.01679111189312405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,768,0.016531555189026725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,12288,0.3399679925706651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,2048,0.014817777607176038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,6144,0.04968888892067803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,32,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,256,0.003519111209445529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,7168,0.0103502223889033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,10240,0.007003555695215861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,64,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,512,0.006775110960006714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,6144,0.02092355489730835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,1024,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,10240,0.014031999640994601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,5120,0.09566755427254571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,2560,0.008984000318580205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,3584,0.0069297779765393995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,3072,0.023060444328520033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,4096,0.011304000185595619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,8192,0.029988444513744775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,7168,0.039020445611741804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,1024,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,65536,0.13503466712103948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,1024,0.019810666640599568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,12288,0.039354665411843195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,768,0.009358222285906473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,256,0.005199999858935674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,128,0.006962666908899943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,2560,0.012456888953844706
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,256,0.03226044442918565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,16384,0.044028444422615894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,2048,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,2560,0.019647111495335896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,1536,0.005974222388532426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,128,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,8192,0.021822222405009802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,512,0.015448000695970325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,4096,0.06862222486072116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,256,0.016559110747443307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,10240,0.016963554753197562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,1536,0.006592888798978593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,4096,0.011397333608733283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,1024,0.006291555447710886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,6144,0.020296888218985665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,10240,0.026033777329656813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,10240,0.007238222493065729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,768,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,6144,0.014524444937705994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,10240,0.03446311089727614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,128,0.00553688903649648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,3584,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,6144,0.02130488885773553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,3584,0.030869334936141968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,4096,0.019509333703253005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,16384,10240,0.05337599913279215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,8192,0.0203813330994712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,7168,0.017123555143674214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,12288,0.35756709840562606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,256,0.025868443979157343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,256,0.011914666328165265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,16384,0.14257155524359807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,12288,0.03537688983811273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,1024,0.007988444632954067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,64,0.009673777553770278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,3072,0.04381688767009311
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,1024,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,16384,0.04288622074657016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,8192,0.010671999719407825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,16384,1024,0.07625155316458808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,5120,5120,0.1158977746963501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,2560,0.01443288889196184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,2560,0.029913776450686987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,16384,0.03221333358022902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,65536,0.10193333360883926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,1536,0.0074275558193524676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,256,0.0035146665241983202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,1536,0.00452622233165635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,1024,0.019108444452285767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,1536,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,7168,0.006636444479227066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,64,0.0472302238146464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,16384,0.03250666790538364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,768,2560,0.007066667079925537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,768,0.006355555521117316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,6144,0.019918221566412184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,16384,0.015460444821251763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,65536,0.01443822185198466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,5120,0.013651555611027611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,1536,0.006340444501903322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,16384,0.09700800312889947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,256,0.003125333330697484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,512,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,12288,0.10719288720024957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,1536,0.010718222293588849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,2560,0.03278577658865187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,10240,0.01420088940196567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,16384,0.016176000237464905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,12288,0.06301955382029216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,2048,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,512,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,512,0.007145778172545963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,16384,0.0322746667597029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,5120,0.03675377700063918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,12288,0.017095999585257638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,32,0.050260444482167564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,512,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,512,0.04236800140804715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,3584,0.013919111755159167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,10240,0.015209777487648858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,3072,0.006275555739800136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,4096,0.007686222592989604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,3584,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,10240,0.11687110530005561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,128,0.0038897775941424896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,65536,0.061944888697730176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,10240,0.011752888560295105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,1024,0.004249777644872665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,7168,0.010701333483060202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,128,0.01001422190003925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,7168,0.010738666686746808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,1536,0.01720711092154185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,4096,0.12204355663723415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,6144,0.015559110376569959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,2048,0.007547555698288812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,3072,0.005575110928879843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3584,128,0.0053075556125905775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,7168,0.02773422168360816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,16384,0.06642133659786649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,16384,0.010020444790522257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,5120,0.009552000297440423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,256,0.0028888889484935333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,12288,0.08964355786641438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,4096,0.008351999852392409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,4096,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,3584,0.007641777396202087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,10240,0.19077777862548828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,7168,0.01204622205760744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,8192,0.1340471108754476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,10240,0.007284444239404466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,1024,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,1024,0.005873777800136142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,768,0.006320888797442119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,16384,0.046527110868030123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,1024,0.015133332875039844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,7168,0.03789510991838243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,65536,0.12362399366166855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,1024,0.005947555518812603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,768,0.008922666311264038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,7168,0.010703999963071613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,65536,4.579475402832031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,256,0.03252266512976752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,7168,0.025898666845427618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,512,0.017197334104114108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,256,0.011349333657158745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,3072,0.025489777326583862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,2560,0.01163822246922387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,16384,0.03711822297837999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,2048,0.015323556131786771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,12288,0.02980888883272807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,3072,0.01000711073478063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,4096,0.007691555553012424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,512,0.005629333357016246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,3072,0.007239110767841339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,65536,0.06782755586836073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,12288,0.015535111228624979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,16384,0.015490666031837463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,3072,0.007243555453088548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,3072,0.014137778017255994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,3584,0.03668533431159125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,256,0.005978666659858491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1024,768,0.015846222639083862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,4096,0.00794666674402025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,8192,0.017823111679818895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,12288,0.04141955574353536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,256,0.005537777725193236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,2048,0.014471999473041959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,5120,0.014450665977266101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,6144,0.013805333111021252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,5120,0.02199911077817281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,65536,0.06677333513895671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,6144,256,0.005888888819350137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,8192,0.010673777924643623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,128,0.014781332678265043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,3584,0.054029332266913525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,3072,0.013453333742088742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,16384,0.008031110796663497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,10240,0.05208177698983086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,12288,0.029322667254341975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,6144,0.03210577699873183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,2048,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,2048,0.08490044540829128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,512,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,7168,0.009666666388511658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,128,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,16384,0.021240888370407954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,256,0.012105777859687805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,3072,0.034838222795062594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,512,0.00490399988161193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,10240,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,768,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,3584,0.007990222010347579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,6144,6144,0.03607644306288825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,12288,0.08048533068762885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,7168,0.01346666696998808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,3072,0.017215111189418368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,3584,0.08356177806854248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,3584,0.007632888853549957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,256,0.0032515554792351196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,256,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,65536,0.04794222116470337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,768,0.015600888265503777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,2560,0.011061333119869232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,8192,0.01590755581855774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,512,0.0052791109515561005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,1536,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,8192,0.02306488818592495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,8192,0.026164443956481084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,6144,0.20062843958536783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,3584,0.011745777395036487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,256,0.0034888887570963967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2560,128,0.005888000130653381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,3584,0.06470844480726454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,65536,0.2747280067867703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,4096,0.025880888104438782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,256,0.005582222094138463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,6144,0.011385777758227454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,768,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,65536,0.20542311668395996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,3072,0.031277332040998675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,256,0.0028986665937635633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,8192,0.01926577753490872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,4096,0.01516888870133294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,7168,0.013095999757448832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,256,768,0.005897777775923411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,2048,0.015172445111804538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,5120,0.015801777442296345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,6144,0.08548355764812893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,512,0.0042035554846127825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,10240,0.11712533897823757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,512,0.0069413334131240845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3584,3072,0.008279999925030602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,10240,7168,0.03353066576851739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,8192,0.03288177649180094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,65536,0.06853866577148438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,6144,0.012096000214417776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,256,0.01683644453684489
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,3584,0.022350221872329712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,5120,0.018522666560279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,2560,0.007063999772071838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,16384,1.2051351335313585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,64,0.03260711166593764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,128,0.0038897775941424896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,1024,0.012072000238630505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,3584,0.012396444049146442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,1536,0.00997866690158844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,6144,0.019111999207072787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,4096,0.00812000036239624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,3584,0.01962222158908844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,8192,0.028322666883468628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,4096,0.01977244516213735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,8192,0.018527110417683918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,128,0.0052151112920708126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,256,0.0076657773719893555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,10240,0.011756444142924415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,7168,0.021448888712459143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,2048,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,512,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,6144,0.022130666507614985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,2048,0.015173332558737861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,32,0.005920889063013925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,10240,0.04404711061053806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,512,0.003181333343187968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,8192,0.01516355574131012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,10240,4096,0.02576622201336755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,6144,0.049624890089035034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,1024,0.016015110744370353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,16384,0.022994667291641235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,7168,0.015481778317027621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,32,0.00794044468137953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,6144,0.009080000221729279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,8192,0.022687110635969374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,8192,0.03020177616013421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,8192,0.011353777514563667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,64,0.002514666567246119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,4096,0.00793866647614373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,7168,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,8192,0.020185778538386028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,3072,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,7168,0.011359111302428775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,768,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,1536,0.07042044401168823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,7168,0.020485333270496793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,1024,0.012534221841229333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,2048,0.006664000037643645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,128,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,8192,0.020745777421527438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,128,0.003096888876623578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,12288,0.016238222519556682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,512,0.00831288927131229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,128,0.005535999933878581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,4096,0.007981333467695449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,768,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,768,0.011071110765139261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,32,0.00554933316177792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,8192,0.0107360002067354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,7168,0.09726666741900974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,512,0.0035342222286595237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,1024,0.006010666489601135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,8192,0.0390035543176863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,5120,0.0927128858036465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,512,0.0038577777643998465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,32,0.007651555869314406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,10240,0.037068443165885076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,65536,0.3130302164289686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,16384,16384,0.09264800283643936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,6144,0.03700177868207296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,7168,0.023944889505704243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,4096,0.05367822117275662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,12288,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,512,0.020413332515292697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,768,0.007547555698288812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,768,0.007250666618347168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,6144,0.008845333423879411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,7168,0.13311110602484807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,12288,0.02397955622937944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,12288,0.1318648921118842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,2560,0.05447377761205038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,12288,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,6144,0.13085777229732937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,768,0.011014222270912595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,128,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,10240,0.015803555647532146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,2048,0.011221333510345884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,5120,0.01611911091539595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,7168,0.014915555715560913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,2048,0.041738665766186185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,1536,0.005050666630268097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,1536,0.017550221747822232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,12288,0.01928444372283088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,65536,0.13511821958753797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,256,0.005384888913896348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,2560,0.035621331797705755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,2048,0.04580444428655836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,12288,256,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,7168,0.08440444204542373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,8192,0.04923199945025974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,65536,0.026974222726292078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,8192,0.0288426677385966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,65536,0.22502843538920084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,256,0.0035128887328836652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,768,0.005699555493063397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,1024,0.016499555773205228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,4096,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,1024,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,7168,7168,0.020417778028382193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,8192,0.03112444281578064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,2048,0.007295110987292395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,1536,0.006454222318198945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,3584,0.017651556266678702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,128,0.01920000049802992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,2048,0.016566221912701923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1536,12288,0.028022223048739966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,3584,0.02407288882467482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,16384,1.1310453414916992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1536,8192,0.04553955462243822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,6144,0.008877333667543199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,2048,0.019856888386938307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,3584,0.011031111081441244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,12288,0.016529776983790927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,4096,0.00755466686354743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,3072,0.007374222079912822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,2048,0.014383999837769402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,32,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,65536,0.12705777751074895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,2048,0.010033778018421596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,3584,0.013218666944238873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,7168,0.07640710804197523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,768,0.01685333251953125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,3072,0.01182044463025199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,2560,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,3584,0.01204177737236023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,128,6144,0.01716444392999013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,16384,0.016184889607959323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,7168,0.03259377678235372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,1536,0.007961778177155389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,5120,0.024357333779335022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,16384,0.03626666797531976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1536,4096,0.011514667007658215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,7168,0.7256426811218262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,12288,0.04186400108867221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,6144,0.007947555846638149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,512,0.020233778489960563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,4096,0.028248889578713313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,16384,0.06051911248101128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,2560,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,5120,0.00868888861603207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,10240,0.03819377885924445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,12288,0.022867555419603985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,256,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,4096,0.00904266701804267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,6144,0.17307555675506592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,2048,0.015895111693276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,6144,0.009276444713274637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,65536,0.07188089026345147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,3072,0.02237155536810557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,1536,0.006311111152172089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,2048,0.022110222114457026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,7168,0.017536888519922893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,512,0.006594666590293248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,2560,0.0075191110372543335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,3584,0.007584888901975419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,8192,2560,0.008970666262838576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,10240,0.030601776308483545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,128,0.011043555206722684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,5120,0.02146133283774058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,7168,0.010683555569913654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,64,0.00796622203456031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,12288,0.042783111333847046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,4096,0.024540444215138752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,2560,0.015555555621782938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,16384,0.040877332290013633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,8192,0.6980346573723687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,3584,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,512,0.013457777599493662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,5120,0.02852355440457662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,128,0.0052675555149714155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,128,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,8192,0.01663644446267022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,5120,0.02474222249454922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,5120,0.021758221917682227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,5120,0.05731111102634006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,5120,0.006632888896597757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,8192,0.018387556076049805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,256,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,4096,0.1517066690656874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,1024,0.0038977778620190094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,3072,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,6144,0.15498133500417074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,512,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,1536,0.006646222124497096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,1024,0.0038817777401871155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,5120,0.009372444616423713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,12288,0.017857778403494094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,768,0.06240622202555338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,256,0.013553777502642738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3584,512,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,6144,0.046925332811143666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,128,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,8192,0.09139733182059394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,16384,0.02645066711637709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,12288,0.07237422466278076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,7168,0.019355555375417072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,10240,0.10742844475640191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,8192,0.03597777750757005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,12288,0.7627315521240234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,128,0.005620444400442972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,1024,0.016240888171725802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,128,0.002887111157178879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,10240,0.023195554812749226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,128,0.003216888962520493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,256,0.015487111277050443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,2048,0.008661333057615492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,10240,0.02951822347111172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,128,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,512,0.020661332541041903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,65536,0.061568889353010386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,1536,0.013821333646774292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,16384,0.020255110330051847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,3072,0.007589333587222629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,3584,0.044621331824196704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,65536,0.1144275532828437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,1536,0.011924444801277585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,128,0.002806222273243798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,16384,1024,0.024756444825066462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,512,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,64,0.005681777579916849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,2048,0.1605786614947849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,5120,0.009416888985368941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,7168,0.014836443795098199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,768,0.026307556364271376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,3584,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,1024,0.013231111069520315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,3072,0.025799110531806946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,32,0.00665866666369968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,7168,0.02362488872475094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,10240,0.01369066701995002
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,3072,0.018765333626005385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,32,0.003944000022278892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3072,256,0.01682844426896837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,6144,0.06296888987223308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,12288,0.07201244433720906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,12288,0.011234666738245221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,2560,0.006688888702127669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,12288,0.01664177742269304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,128,0.00300177786913183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,512,0.007992888490358988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,10240,0.014467555615637036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,65536,0.02887111239963108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1024,7168,0.07730844285753039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,1024,0.01723288827472263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,512,0.00423911131090588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,1536,0.012097777591811286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,6144,0.012347555822796293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,768,0.02260711126857334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,12288,0.016943999462657504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,2560,0.006161777923504512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,1024,0.005976000179847081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3072,128,0.014487110906177096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,12288,0.014104889498816596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,3584,0.02093066606256697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,768,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,3072,0.00832088871134652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,3072,0.00775999989774492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,7168,0.008740444150235917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,16384,0.01796266602145301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,2048,0.013880000346236758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,4096,0.010082667072614035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,512,0.015763555963834126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,2560,0.013353777428468069
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,2048,0.029156443145540025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,8192,0.01613066593805949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,2048,0.006665777828958299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,512,0.007063111497296228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,16384,0.020672889219390023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,256,0.0136817776494556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1536,1024,0.0042275554604000514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,5120,0.008319999608728621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,2560,0.017247110605239868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,256,0.0052666668262746595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,6144,0.019550222489568923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,1024,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,4096,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,64,0.0107004443804423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,10240,0.016144000821643405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,3072,0.007320889168315464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,128,0.00314666661951277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,1024,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,3072,0.013535111314720578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,128,0.007227555745177799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,1024,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,7168,0.0935022234916687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,6144,0.05048622356520759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,3072,0.0063662222690052455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,3584,0.0075777777367168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,3072,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,5120,0.02390577726893955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,7168,0.010384000009960597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,768,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,768,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,128,0.0038622220357259116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,64,0.0034808889031410217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,5120,0.027308444182078045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,128,0.0028897778441508612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,256,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,512,0.013820444544156393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,5120,0.05879555808173286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,128,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,12288,0.019367110398080613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,3072,0.015106666419241162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,2560,0.007297777467303806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,2560,0.007088888850477006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,65536,0.03159911102718777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,4096,0.11043200227949355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,32,12288,0.008669333325492011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,3584,0.032256888018714056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,65536,0.236791107389662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,8192,0.013160000244776407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,2560,0.047353777620527476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,3072,0.0090506664580769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,512,0.015549333559142219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,768,0.005578666511509154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,7168,256,0.005576888720194499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,16384,0.016164445214801364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,8192,0.014581332604090372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,3072,0.005913777897755305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,5120,0.030692444907294378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,5120,0.06553600231806438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,128,0.011074666347768573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,768,10240,0.012480888929631976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,2048,0.2998631000518799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,12288,5120,0.18535821967654756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,64,0.009569777382744683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,2048,0.03641688823699951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,512,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,16384,0.04172711239920723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,1536,0.030028445853127375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,65536,32,0.012159110771285163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,8192,0.01072088877360026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,1536,0.004905777672926585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,128,0.005274666680230035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,256,0.006970666348934174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,5120,1.3090231153700087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,2048,0.011718221836619906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,10240,0.017921777235137094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,7168,0.02628355556064182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,32,0.0028159999185138275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,768,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,65536,0.11134666866726345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,2048,0.009699555734793345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,12288,0.0355404449833764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,256,0.0034968890249729156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,1536,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,1536,0.004857777721352047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,3072,0.0844586690266927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,32,0.025518221987618342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,8192,0.030596445004145306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,256,256,0.005256888767083486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,512,0.014845333165592618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,2048,0.01738666660255856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,768,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,1024,0.003910222401221593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,5120,0.027493332823117573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,2048,0.006430222342411677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,768,0.005997333261701796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,1024,0.02141511109140184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,5120,0.021213332811991375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,3584,0.007239999870459239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,5120,0.008933333059151968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,12288,0.03659910957018534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,12288,0.03642488850487603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,1536,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,64,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,3584,0.007341333561473423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,1024,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,512,0.005623111294375525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,8192,0.03959644503063626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,65536,16384,0.3067306677500407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,65536,0.19272889031304252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,2560,0.007963555554548899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,7168,0.26987822850545246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,3584,0.019546666079097323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,1024,0.02734222180313534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,5120,0.028746667835447524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,4096,0.02564711040920681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,128,0.017364444004164804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,8192,0.06427022483613756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,6144,1536,0.015825778245925903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,4096,0.011678222152921887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,3072,0.0310444434483846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,3584,0.007609777980380588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,6144,1536,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,512,0.003549333247873518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,128,0.003257777748836411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,12288,0.045251554912990995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,4096,0.008277333445019191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,256,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,65536,0.009024889104896123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,768,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,1024,0.019899555378490023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,128,0.004464000049564573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,10240,0.011159111228254108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,65536,0.2281066576639811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,2048,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,512,0.0052560000783867305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,128,0.005565333283609814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,2560,0.045399109522501625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,12288,0.25978488392300075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,65536,0.0596675541665819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,12288,0.014125333891974555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,3584,0.036277333895365395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,12288,0.013938667045699226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,6144,0.09593155648973252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,4096,0.022191999687088862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,10240,0.0395395557085673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,2048,0.026208889153268602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,512,0.021012443635198805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,64,0.00675999994079272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,7168,0.05938755803638034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,7168,128,0.009080000221729279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,1536,0.008776889079146916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,1024,0.08382933007346259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,32,0.008025777836640676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,2560,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,3072,0.027440889014138117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,2560,2.846861309475369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,10240,0.012446222205956778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1024,3584,0.013422222601042854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,256,0.0062871111763848206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,5120,0.026781333817376032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,2048,0.0270106659995185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,1024,0.0058542220956749385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,128,0.003543111185232798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,12288,0.06790844599405925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,4096,0.008100444244013892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,3072,0.011024889018800525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,128,0.008388444781303406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,1536,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,6144,0.013207111093733044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,7168,0.02312444481584761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,768,5120,0.020970667401949566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,1536,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,1024,0.016926222377353244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,3072,0.01309333327743742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,3584,0.008301333420806462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,5120,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,1536,0.025948445002237957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,4096,0.007965332931942409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,2048,0.009659555223253038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,5120,0.011494222614500256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,1024,0.0143857780430052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,768,0.005984000033802456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,3072,0.010377777947319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,3584,0.007644444704055786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,1536,0.006552000012662675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,2560,0.00701422244310379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,65536,0.1497466696633233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,2560,0.0069324444565508105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,768,0.40817422337002224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,10240,0.030831111801995173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,1536,0.05006666647063362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,512,0.006960888703664144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,16384,0.05766044722663032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,65536,0.09042488866382176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,65536,0.06481599807739258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,256,0.004550222307443619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,1536,0.014146667387750415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,7168,0.02599999970859951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,1536,0.014420444766680399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,3584,0.01070666644308302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,8192,0.034268445438808866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,4096,0.030113776524861652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,4096,0.00832799987660514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,12288,0.02437244521247016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,8192,0.014816888504558139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,2048,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,4096,0.029312888781229656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,4096,0.037963555918799505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,64,0.003195555466744635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,10240,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,1536,0.019539554913838703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,1536,0.017860444055663217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,32,0.005879999862776862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,64,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,768,0.005666666560702854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,128,0.00798044436507755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,256,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,8192,0.0314284430609809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,10240,0.016229333149062265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,4096,0.10563288794623481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,5120,0.01204177737236023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,16384,0.16323199537065294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,256,64,0.0028204443968004654
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,2048,0.0069297779765393995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,16384,0.02089333368672265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,3072,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,768,0.030121776792738173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,2560,0.007243555453088548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,2048,0.005192888693677055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,3072,0.020830222302012973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,64,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,3584,0.04860088891453213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,2048,0.021205332544114854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,16384,0.01864266726705763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,512,0.01959822244114346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,6144,0.06689511405097114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,2560,0.16330933570861816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,2048,0.006948444578382704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,4096,0.06744177473915948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,16384,0.01973777843846215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,256,0.0033200000309281876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,1024,0.004036444342798657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,1024,0.0074986666440963745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,3584,0.009185777770148383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,10240,0.019366222951147292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,6144,0.012488889197508493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,5120,0.010670222342014313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,2560,0.03292355603641934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,1536,0.010760889285140567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,128,0.0028080000645584534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,1024,0.015110222829712762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,512,0.005933333188295364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,3584,0.01748533381356133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,8192,0.3033191098107232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,1536,0.008887111312813228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,12288,0.016384000579516094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,6144,0.017453332742055256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,3072,0.00555288874440723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,65536,0.11026310920715332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,2560,0.03144977821244134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,128,0.010801778071456485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,64,0.003891555385457145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,65536,0.1488800048828125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,256,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,3584,0.008948444492287105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,1536,0.011359111302428775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,32,0.005719999886221356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,8192,0.008965333302815756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,512,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,2048,0.00977244476477305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,3584,0.009676444033781687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,2560,0.015152000718646579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,1536,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,32,0.006235555642180973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,10240,0.016876444220542908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,2048,0.06620355447133382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,5120,0.03222489025857713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,65536,0.14506044652726915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,16384,0.026908444033728704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,5120,0.018224888377719455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,1024,0.02316444449954563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,768,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,1024,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,128,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2048,10240,0.026897778113683064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,8192,0.042154666450288564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,3584,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,768,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,5120,0.01700622174474928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,3072,0.036601778533723615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,1536,0.016560888952679105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,768,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,256,0.014811555544535318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,128,0.0041955556306574075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,65536,0.07935733265346952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,12288,64,0.01961955593691932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,128,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,128,0.011838222543398539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,4096,0.016347555650605094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,2048,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3584,8192,0.024714666936132643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,512,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,12288,0.029535998900731403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,2560,0.03127288818359375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,16384,0.8714950879414877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,4096,0.05363022287686666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,5120,0.011405333048767514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,12288,0.020456888609462313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,3072,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,3584,0.024400000770886738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,128,0.012732444538010491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,12288,0.013438222308953604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,512,0.0038319999972979226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,7168,0.02129777769247691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,12288,0.016579555140601266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,128,0.0052151112920708126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,12288,0.02702577743265364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,65536,1.113200929429796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,7168,0.019917334119478863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,256,0.004556444370084339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,7168,0.04036444425582886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,1024,0.03245333168241713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,6144,0.027518222729365032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,512,0.004872000051869286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,7168,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,2560,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,512,0.006242666807439592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,65536,0.07796266343858507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,64,0.016524444023768108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,2560,0.006581333362393909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,12288,0.13065244091881648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,16384,0.04342755675315857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,3072,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,12288,0.019235556324323017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,16384,0.019640889432695176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,6144,0.11258578300476074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,4096,0.017660443981488545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,65536,1.2744017706976998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,2048,0.011143999795118967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,16384,0.048639999495612256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,2048,0.009409777820110321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,4096,0.014879110786649915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,12288,0.087117334206899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,4096,0.007671111159854465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,2560,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,7168,0.02239377796649933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,2560,0.009367110828558603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,512,0.02197155521975623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,1024,0.00423555572827657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,16384,0.37951021724277073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,128,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,32,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,2048,0.007299555672539606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,128,0.003126222226354811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,2048,0.01961599952644772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,512,0.01514311134815216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,65536,2.6211456722683377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,128,0.004953777624501122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,6144,0.011343110766675739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,5120,0.012513777448071374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,768,0.0038693332009845306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,1536,0.004860444615284602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,1024,0.08039644691679212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2048,2048,0.01519377695189582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,128,0.05453866720199585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,12288,1024,0.00941955546538035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,1536,0.007958222594526079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,1536,0.016220443778567843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,1536,0.00833422193924586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1024,8192,0.07506310939788818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,32,0.007924444145626491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,512,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,512,0.013033777475357056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,1024,0.006188444379303191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,256,0.015169777803950839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,65536,0.13622132937113443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,7168,0.015793777174419828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,12288,0.21967821651034883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,768,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,4096,0.007972444097201029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,256,3584,0.00699733363257514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,12288,0.03742400142881606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,256,0.016864000095261466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,1024,0.016229333149062265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,256,0.0035013332962989807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,3072,0.031217777066760596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,1536,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,512,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,5120,0.009791110952695211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,3584,0.005607999861240387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,5120,0.020592000749376085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,1024,0.0052622221410274506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,768,0.13607466220855713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,3072,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,1536,0.01276266657643848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,2048,0.006996444529957241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,1024,0.01514311134815216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,3584,0.08946400218539768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,32,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,768,0.01551466683546702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,32,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,16384,0.06403377983305189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,65536,0.35258844163682723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,2560,0.00665511108107037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,3584,0.02242311173015171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,12288,0.263084438112047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,12288,0.032469332218170166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,8192,0.0079644446571668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,3072,0.007694222033023834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,512,0.0042160000238153665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,4096,0.007972444097201029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,10240,0.051721778180864125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,8192,0.09327822261386448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,16384,0.026153778036435444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,256,0.01514400045077006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,512,0.0031600000543726813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,6144,0.1538026730219523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,10240,0.203219559457567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,768,0.011592888997660743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,5120,0.01629866659641266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,5120,0.10363111231062148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,6144,0.01237422227859497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,128,0.037321776151657104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,768,0.004210666649871402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,512,0.010384000009960597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,10240,0.016660445266299777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,5120,0.013793778088357715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,3584,0.010400888820489248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,768,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,7168,0.022448889083332483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,512,0.01679111189312405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,16384,0.020524443851576913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,768,0.025506666964954797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,256,0.0028782222006056043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,768,64,0.0028239999794297745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,2048,0.012290666500727335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,5120,0.021648888786633808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,768,0.024545777175161574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,4096,0.00675644435816341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,7168,1536,0.006799110935793982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,5120,0.010564444793595208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,5120,0.03942844602796767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,768,0.017495110630989075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,1024,0.0151573336786694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,4096,6144,0.01351288871632682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,2048,0.01072177787621816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,6144,0.0090515555606948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,6144,0.011216888825098673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,128,0.01442488862408532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,7168,0.03373510970009698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,2560,0.017996443642510306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,10240,0.02460711035463545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,12288,3.2283751169840493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,3584,0.008729777402347988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,3072,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,4096,1536,0.006304889089531369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,12288,4096,0.014429334137174817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,512,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,65536,0.04527555571662056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,16384,0.07845866680145264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,10240,0.10613066620296902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,8192,0.01479644411140018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,2560,0.009023110899660323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,512,0.00638133328821924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,6144,0.025623111261261836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,5120,0.005260444349712796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,16384,0.05206222335497538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,2560,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,512,0.016148444679048326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,16384,0.015867556134859722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,3072,0.03956177830696106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,2560,0.021715554926130507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,1536,0.008752889103359645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,128,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,2560,0.006100444330109491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,5120,5120,0.03151822090148926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,6144,0.0959635575612386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,3584,0.03828000028928121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,1536,0.029725333054860432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,65536,0.055211553970972695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,3072,0.015136000182893542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,32,0.0028497777465316984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,32,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,16384,0.1048426694340176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,2048,0.02349600030316247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,1024,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,4096,0.010377777947319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,3584,0.009698666632175446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,6144,0.01682933337158627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,1536,0.00794222205877304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,7168,0.0417520006497701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,8192,1536,0.024668445189793903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,16384,0.0332640012105306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,12288,0.0217866665787167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,16384,0.025662221842341956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,16384,0.033562666840023465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,2560,0.007240888973077138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,64,0.004576888763242298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,32,0.006215111249023014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,16384,0.06891466511620416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,65536,0.11826577451494004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,5120,0.009703111317422655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,8192,0.008284444610277811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,12288,0.025849777791235182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,2048,0.007028444773621029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,2560,0.014712888333532544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,1024,0.029892444610595703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,16384,2560,0.01628800067636702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,128,0.014156444205178155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,16384,0.06967200173272027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,1024,0.009358222285906473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,16384,0.03671733207172818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,5120,0.027095999982621934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,128,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1024,512,0.008955555657545725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,5120,0.01074044406414032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,7168,0.023448000351587932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,2048,0.005993777679072485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,12288,0.02538577715555827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,3584,0.014589332871966891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,128,0.005528888768619961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,16384,0.01610311037964291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,8192,0.014436443646748861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,65536,0.5144533581203884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,1024,0.030650665362675984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,1536,0.005183111048407025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,3072,0.007234666910436418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,5120,0.010090666512648264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,2560,0.07386666536331177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,2560,0.013447110851605734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,6144,0.01793333391348521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,768,0.015861334072219003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,16384,0.028836445675955877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,256,0.0052533331844541765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,2048,0.04695733388264974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,256,0.01551822159025404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,10240,0.08347288767496745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,3072,0.0069982219073507525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,5120,0.024808888634045918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,512,0.015124445160230001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,10240,0.012426666915416718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,1024,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,128,0.01480977733929952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,7168,4096,0.012759110993809171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,4096,0.0076453329788313965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,1024,0.0038284444146686127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,65536,0.1588160064485338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,2048,0.01576266686121623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,3072,0.03647733396954007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,128,0.005967999911970562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,256,0.004536889079544279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,8192,2048,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,512,0.023077333966890972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,65536,0.1314444409476386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,1536,0.13602756129370794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,1024,0.026857778429985046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,128,0.007012444237867991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,16384,0.024878222081396315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,3584,0.13392089472876653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,128,0.005208888815508948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,16384,0.3448595470852322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,3072,0.00703288863102595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,256,0.0123857781291008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,3072,0.014615999327765571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,12288,0.03698666559325324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,4096,0.03154844376775954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,65536,0.1924097802903917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,16384,0.02556266718440586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,512,0.0032017777363459268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,12288,0.014108444253603617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,32,0.003170666595300039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,3072,0.01665422154797448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,5120,0.021617778473430212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,32,0.003156444471743372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,64,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,256,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,2048,0.015153777268197803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,6144,0.6586204634772407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,12288,0.048839999569786914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,6144,2560,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,4096,0.029098666376537744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,5120,0.011393778026103973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,8192,0.03771200113826328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,2048,0.4311982260810004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,256,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,1536,0.011063999599880643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,768,0.007915555602974361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,16384,0.02275466587808397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,3072,0.007304888632562425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,2560,0.023044443792766992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,7168,0.027793778313530818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,7168,0.10755199856228298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,6144,0.02130222154988183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,3072,0.016529776983790927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,3584,0.01238044434123569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,16384,0.2844382127126058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,2048,0.1563911173078749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,512,0.007375111182530721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,1536,0.00904799997806549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,64,0.00998399986161126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,768,0.021648888786633808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,10240,0.023023999399609033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,64,0.007937777373525832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,5120,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,6144,0.032814221249686346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,16384,0.023040889037979975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,16384,0.016936888297398884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,3072,0.007643555601437886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,3072,0.017439999514155917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,256,0.015586665934986539
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,3072,0.02308888898955451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,512,0.003906666818592283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,2048,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,4096,0.016160888804329764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2048,3072,0.006467555546098285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,65536,0.08103733592563205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,3072,0.021894221504529316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,3072,0.03888533181614346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,7168,0.012044444680213928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,10240,0.035863998863432146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,3072,0.06905955738491483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,128,0.0028897778441508612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,12288,0.02049688829316033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,1536,7168,0.01260444439119763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,256,0.008636444807052612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,8192,0.3226666715410021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,7168,16384,1.058367093404134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,3072,0.03359377715322707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,768,0.008310221963458592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,16384,0.1111182239320543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,12288,1.5213111241658528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,3584,0.01349422252840466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,3584,0.017856000198258292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,256,0.015483554866578845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,65536,0.1742933326297336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,7168,0.043771555026372276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,5120,0.02914666798379686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,10240,0.009311999711725447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,12288,0.013651555611027611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,512,0.01253600004646513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,10240,0.09993955824110244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,3584,0.029665778080622356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,1536,0.0158942225906584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,4096,0.032948444286982216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,8192,0.011445333560307821
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,65536,0.06314222017923991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,3072,0.007270221908887227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,768,0.051623112625545926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,3072,0.01348888874053955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,7168,0.013762666947311826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,2560,0.020974222156736586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,768,1024,0.01332266628742218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,7168,0.025979555315441553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,16384,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,2048,0.02885777751604716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,512,0.003536000019974179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,768,0.01907111042075687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,8192,0.01069155583779017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,2048,0.0052639999323421055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,6144,8192,0.18148977226681182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,128,0.013765333427323235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,2048,0.0063475556671619415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,7168,0.008303999900817871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,16384,0.31811287668016225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,65536,0.0551546679602729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,4096,0.04876177840762668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,10240,12288,0.03353066576851739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,65536,7168,0.12579466236962214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3072,12288,0.027618666489919026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,512,0.003177777760558658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,128,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,10240,0.028210666444566514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,65536,0.17833688524034289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,5120,0.010260444548394945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,4096,0.01697688963678148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,6144,8192,0.023185777995321486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,7168,0.010135110881593492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,128,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,768,0.028929776615566675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,128,0.014771555860837301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,7168,0.010363555616802638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,16384,0.08296710915035672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,5120,0.018974221414989896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,10240,0.06946933269500732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,2048,0.0181368887424469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,5120,0.012367111113336353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,64,0.010294222169452244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,65536,0.08396177821689182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,1536,0.05986489189995659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,64,0.011496889094511667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,256,0.003157333367400699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,4096,0.01824355622132619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,2560,0.006991110742092133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,8192,0.20153066847059461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,256,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,3584,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,512,0.0063795554969045855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,4096,0.009946666657924652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,512,0.005618666609128316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,768,0.018292443619834054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,2560,0.06291911337110731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,128,65536,0.015224888920783997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,3072,0.008036444584528605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,12288,0.02404266595840454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,768,0.04197866717974345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,768,0.006582222051090664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,768,0.0038417776425679526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,16384,0.015501333607567681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,12288,0.01311555587583118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,8192,0.01719911065366533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,16384,0.03449511196878221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,5120,12288,0.19942933983272979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,10240,3584,0.014375110467274984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,65536,0.011706666813956367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,12288,0.029101332028706867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,256,0.00351200004418691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,768,0.003657777690225177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,1536,0.004579555657174853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,7168,0.13062311543358698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,3584,0.022063111265500385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,256,0.006664000037643645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,3584,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,1536,0.015234667393896313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,65536,0.06163022253248426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,7168,0.05032799972428215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,3584,0.008650666309727563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,32,0.00282222218811512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,32,0.007631111476156447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,7168,0.03567377726236979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,10240,0.1354622178607517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,2560,0.022537777821222942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,5120,0.010025777750545079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,7168,0.1467448870340983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,16384,0.13126577271355525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,6144,8192,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,4096,0.02028888960679372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,3072,0.018165333403481376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,10240,0.01885688967174954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,3072,0.018796443939208984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,3072,0.007622222105662028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,32,0.002790222151411904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,7168,0.019205333458052743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,7168,0.016207999653286405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,128,0.0052693333062860705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,3072,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,5120,0.010407999985747868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,1536,3072,0.01647911138004727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,10240,16384,0.06609155734380086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,512,0.025758221745491028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,2560,0.04234133495224846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,4096,0.019690665933820937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,5120,0.030974222554100886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,4096,0.014670222997665405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,1024,0.03749333487616645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,2048,0.02719199988577101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,256,0.037111110157436795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,1536,0.05304000112745497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,6144,0.01757600075668759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,65536,0.06311021910773383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,32,0.0032035555276605818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,512,0.005921777751710679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,3584,0.008672888908121321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,3072,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,1024,0.004898666507667965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,4096,0.00902844468752543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,2560,0.006934222247865465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,256,0.014985778265529208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,2048,0.006331555545330048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,3584,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,12288,0.07019555568695068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,1536,0.008279999925030602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,64,0.003527111063400904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,32,0.04220889012018839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,8192,16384,0.045385777950286865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,4096,0.007986666427718269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,12288,0.16533599959479436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,6144,0.024711110525661047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,4096,0.03586133321126302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,1536,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,768,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,8192,0.08694666624069214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,12288,0.026679999298519556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,512,0.006574222197135289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,2048,0.005542222410440445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,10240,0.019698666201697457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,1536,0.008788444101810455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,256,0.003253333270549774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,1024,0.005872000008821487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,4096,0.1613688866297404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,16384,0.020826665891541373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,3072,0.09979377852545844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,5120,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,512,0.012061333490742577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,4096,768,0.01681600014368693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,5120,0.00851111114025116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,2048,0.006626666833957036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,8192,0.010669333239396414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,768,0.005608888963858287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,1024,0.006329777754015393
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,65536,0.0881120032734341
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,6144,0.010766222245163388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2560,65536,1.6493404176500108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,256,0.006650666809744305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,8192,0.02131822208563487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,128,0.023023110296991136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,128,0.0052879999081293745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,3072,0.015517334143320719
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,512,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,8192,0.022292445103327434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,1024,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,1024,0.014797333214018079
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,2560,0.03791288865937127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2048,512,0.013419555293189155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,3584,0.009152889251708984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,1536,0.013823111024167804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,6144,0.011944889194435544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,16384,0.027950220637851294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,10240,0.011742221812407175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,128,0.005943999936183293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,8192,0.010399999717871347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,6144,0.009646221995353699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,64,12288,0.0075902218619982404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,3072,0.018975110517607797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,3584,0.020530665914217632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,7168,0.023383110761642456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,5120,0.017844445175594754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,8192,0.19965067174699572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,7168,0.012075555821259817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,2560,0.035960889524883695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,3072,0.013530666629473368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3584,65536,0.21563288900587294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,7168,0.051742222574022084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,7168,0.03748977846569485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,7168,0.009712888962692684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,1536,0.0069724445541699724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,5120,0.008671110702885522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,16384,0.16370044814215765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,3072,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,6144,0.017639999588330586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,6144,0.03388799892531501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,12288,0.013177778157922955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,6144,0.14565332730611166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,7168,0.04757066567738851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,2048,0.006797333144479328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,65536,0.0658151110013326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,3072,0.2863351239098443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,768,0.003922666526503033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,6144,0.009166222479608323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,256,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,16384,0.3117920027838813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,2560,0.005921777751710679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,128,0.003216000066863166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,2560,0.014905777242448596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,8192,0.010444444086816577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,3072,0.007394666473070781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,8192,0.02163200080394745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,256,0.014760888285107084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,512,0.0035866668654812705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,8192,0.018982221682866413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,16384,0.026913776993751526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,65536,128,0.018242667118708294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,6144,0.02076088885466258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,512,0.006590222318967183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,6144,0.17874756124284533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,12288,0.10869866609573364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,10240,0.015024888846609326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,1024,0.005937777873542573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,5120,0.012392000191741519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,2048,0.014974221587181091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,512,0.007119110888904995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,16384,0.25094487931993276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,1024,0.004209777961174647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,12288,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,6144,0.015185778339703878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,65536,0.06603199905819364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,10240,0.02625511089960734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,8192,0.06296533346176147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,65536,1.562289767795139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,1024,0.005562666803598404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,64,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2048,4096,0.007630222373538547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,65536,768,0.03822933302985297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,256,0.05195822318394979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,2560,0.006233777850866318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,512,0.003519111209445529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,4096,0.023111111587948267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,10240,5120,0.320066663953993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,16384,0.04534844557444254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,512,0.014128888646761576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,64,0.011643555429246692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,3584,0.02195555633968777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,768,5120,0.018233777748213876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,16384,0.027469333675172593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,7168,0.013789333403110504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,1536,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,3072,0.008762666748629676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,16384,0.02917511264483134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,64,0.0028826666788922418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,5120,0.008834666675991481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,6144,0.018466666340827942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,4096,0.006631999793979857
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,10240,0.037371555964152016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,512,7168,0.00962222201956643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,3072,0.007628444168302748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,7168,0.3281324439578586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,512,0.008664888640244802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,5120,0.01686311099264357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,8192,0.013071999781661563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,5120,0.020301333732075162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,512,0.020985777179400127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,2048,0.033181332879596286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,1536,0.008078222473462423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,10240,0.018967999352349173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,7168,0.034677333301968045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,512,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,8192,0.005893333504597346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,8192,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,10240,0.08416799704233806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,7168,0.01350400017367469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,7168,0.014067555467287699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,5120,0.021219554874632094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,16384,0.046498666206995644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,1536,0.010751999914646149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,128,0.006600889066855113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,7168,0.01069866700304879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,16384,0.04075288772583008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,6144,3.7004606458875866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,65536,0.07281955745485094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,512,0.09786311123106216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,3072,0.013457777599493662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,65536,0.0868026680416531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,1024,0.0636426673995124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3072,128,0.010267555713653564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,768,0.005879999862776862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,8192,0.010704889065689512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,3072,0.007582222421964009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,12288,0.016923555069499545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,512,0.1078195571899414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,256,0.003548444559176763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,1024,0.005595555735958948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,7168,0.012113778127564324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,10240,0.012100444071822695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,3584,0.6762906710306803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,7168,0.01260355528857973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,512,0.005682666682534748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,8192,0.02484888831774394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,128,0.005666666560702854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,3072,0.01625688870747884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,3584,0.014522666732470194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,8192,8192,0.01939733326435089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,5120,0.009767110976907942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,128,0.00629688882165485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,1536,0.00655733338660664
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,7168,0.012545777691735161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,65536,0.07910755607816908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,1536,0.010374222364690568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,512,0.0069271110826068455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,128,0.012242666549152799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,65536,0.06341955396864149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,7168,0.02604355580276913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,64,0.003689777933888965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,64,0.0028444443725877335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,5120,0.008282666405042013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,32,0.020900444851981267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,7168,0.02369600037733714
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,256,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,1536,0.018023111753993563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,16384,0.019891555110613506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,4096,0.04299022091759575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,512,0.0069253332912921906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,256,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,65536,16384,0.16005333264668783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,10240,0.2153866688410441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,6144,0.00960711141427358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,12288,0.013424889081054263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,64,0.003059555465976397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,16384,4.375367058648004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,128,4096,0.011013333168294696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,512,0.014565333724021912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,256,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,7168,0.013824889229403602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,8192,0.015293333265516492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,768,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,8192,0.026015111141734656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,8192,0.06982844405704074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,10240,0.022809776994917128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,8192,0.012274666792816587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,16384,0.05150666501786974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,10240,0.050514668226242065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,32,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,12288,0.033703111939960055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,3072,0.0099973330895106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,12288,0.007931555310885111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,32,0.009709333380063375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,10240,0.0262755552927653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,4096,0.01811644434928894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,1024,0.004247111164861255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,7168,0.00831822223133511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,3072,12288,0.03295733200179206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,10240,0.007329777710967594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,2560,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,7168,0.024725332856178284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,3072,0.03772355450524224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,128,0.02401688860522376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,65536,0.0550871127181583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,2560,0.009019555317031013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,128,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,2048,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,768,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,1536,0.021169778373506334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,4096,0.008681777450773451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,256,0.0031475555151700974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,6144,0.05058311091528999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,2048,0.021182222498787776
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,768,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,5120,0.022669333550665114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,65536,0.32496444384257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,3072,0.0313937763373057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,128,0.005218666460778978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1536,2048,0.007324444750944774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,6144,0.1558204491933187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,2048,0.006945778098371293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,256,0.0034933334423436057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,65536,0.12740177578396267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,2560,0.014382221632533602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,12288,0.02239111065864563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,128,0.005250666704442766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,5120,7168,0.012337778177526263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,7168,0.009668444593747457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,768,0.016879111528396606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,768,0.016859556237856548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,3584,0.007304888632562425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,2048,0.006321777900060018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,4096,0.03369333346684774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,768,0.0069777775141927935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,4096,0.010417777631017896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,2048,0.006457777900828256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,65536,0.3168586624993218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,1024,0.019000889526473153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,3072,0.012077333198653327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,10240,0.02276444435119629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,16384,0.009367110828558603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,1024,0.01515111161602868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,2560,0.015104000767072042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,5120,0.18884621726142037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,64,0.0028773333049482773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,512,0.008593777815500895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,4096,0.00721777809990777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,1536,0.015344889627562629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,3072,0.017847999930381775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,16384,0.04507022102673849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,4096,0.008279110822412703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,65536,0.22835289107428658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,4096,0.008532444636027018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,3584,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,65536,0.06567111280229357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,10240,0.014799111419253878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,1024,0.006190222170617845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,1536,0.00996888925631841
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,6144,0.02772888872358534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,256,0.00926577796538671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,10240,0.02978488802909851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,12288,0.013072888884279462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,65536,0.3539902104271783
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,1024,0.007311111523045435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,2560,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,1024,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,12288,0.08969066540400188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,3584,0.006608888920810487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,1024,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,1536,0.015156444576051501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,2560,0.006984000404675801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,12288,0.011014222270912595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,512,0.008025777836640676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,5120,1536,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,256,0.0052586665583981415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,2560,0.020586666133668687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,5120,0.00628266649113761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,32,0.01034400032626258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,2560,0.015477332803938123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,512,0.01517599986659156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,12288,0.024995555480321247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,32,0.00554577757914861
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,3072,0.030976888206270006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,32,0.02797599964671665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,10240,0.012661333713266583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,1536,0.012444444828563266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,512,0.007310222420427535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,512,0.020430222153663635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,3072,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,6144,0.009135999613338048
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,16384,0.10899022552702163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,65536,0.061866667535569936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,32,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,16384,0.06928355826271905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,6144,0.013100444442696042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,2048,0.016220443778567843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,3072,0.02742133372359806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,3072,0.007596444752481248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,768,0.007035555938879649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,512,0.0163137780295478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,7168,0.028502222564485338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,4096,0.06586133109198676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,4096,0.022830223043759663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,768,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,3072,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,8192,0.023359111613697473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,768,0.003588444242874781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,10240,0.021172444025675457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,1024,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,5120,0.016718222035302054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,2048,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,128,0.00775911079512702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,7168,0.011744000017642975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,7168,0.13194755713144937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,12288,0.020668443706300523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,2048,0.008952889177534316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,4096,0.051615999804602734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,6144,0.01442488862408532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,5120,0.01686044368478987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,7168,0.03253066539764404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,768,0.005968889014588461
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,1536,0.026621333426899377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,2048,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,65536,0.07226488987604777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,2560,0.027525333894623652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,8192,0.19044266806708443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,256,0.018533334136009216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,16384,0.045775112178590566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,128,0.0028293333533737394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,64,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,3584,0.03353510962592231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,3072,0.08385422494676377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,5120,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,768,0.006958222223652734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,3072,0.00740444411834081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,12288,0.038239111502965294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,8192,0.02611911131276025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,16384,0.023623110519515142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,6144,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,65536,0.048784001006020435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,32,256,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,64,0.008280889027648503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,3072,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,32,0.03277955452601115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,12288,0.041067557202445135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1024,2560,0.005931555396980709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,768,0.01613066593805949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,768,0.008169777691364288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,128,0.0058355554938316345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,7168,0.1499155494901869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,1536,0.008656889200210571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,3072,0.007619555625650618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,65536,3.363252427842882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,65536,2048,0.07571999894248115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,5120,0.018788443671332467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,6144,0.012125333150227865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,65536,0.23215911123487684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,6144,0.014150222142537435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,512,0.013793778088357715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,10240,0.1640746725930108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,5120,0.10685066382090251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,768,0.005932444499598608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,3584,0.01516266663869222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,6144,0.010662222074137794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,8192,0.007992888490358988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,1536,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,3072,0.009319110876984065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,6144,0.014186667071448432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2560,5120,0.02269422180122799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,6144,0.009717333647939894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,7168,0.014119111829333834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,64,0.002918222298224767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,16384,0.028592887851927016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,1024,0.013777777552604675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,65536,0.12864532735612658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,10240,0.012058667010731168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,12288,0.04488889045185513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,768,0.003907555507289039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,6144,0.04979022343953451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,6144,0.017662222186724346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,128,0.003826666623353958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,6144,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,768,0.007904888855086433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,768,0.00793599999613232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,128,0.017848889032999676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,1024,0.005903111149867375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,32,0.0028080000645584534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,16384,0.03600800037384033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,8192,0.015543111496501498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,64,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,1024,0.007797333101431529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,128,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,5120,0.017158221867349412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,5120,0.014905777242448596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,7168,0.018840889135996502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,8192,0.03048266636000739
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,2560,0.031009776724709406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,1536,0.004506666627195147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,512,0.02060622142420875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,4096,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,3584,0.006666666517655055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,2560,0.010482666393121084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,4096,0.013947554760509066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,512,0.006968888971540663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,8192,0.02060622142420875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,512,0.013528889252079858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,5120,0.008674666285514832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,768,0.0038088887102074097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,5120,3072,0.025731555289692346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,1024,0.020615110794703167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,3584,0.012959111067983838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,768,0.061864886018964976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,2560,0.00655644428398874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3584,256,0.006240889016124938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,2048,0.009347555538018545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,256,0.003173333282272021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,8192,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,5120,3072,0.008649778034951951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,1024,0.04039377636379666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,3072,0.007625777688291337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,1536,0.046352889802720815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,6144,16384,0.02296888828277588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,6144,0.02053866618209415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,128,0.015133332875039844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,4096,8192,0.025009777810838487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,5120,0.01923822197649214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,6144,0.02454311152299245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,65536,0.21220000584920248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,1024,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,256,0.06660355461968316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,2048,0.0069671107663048645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,32,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,12288,2048,0.013608000344700284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,10240,0.018210666047202218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,256,0.033569776349597506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,3072,0.0317884451813168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,768,0.0069653333889113525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,16384,0.042952001094818115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,2560,0.0069751110341813825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,16384,0.05529333485497368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,16384,0.03086755673090617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,7168,0.013479111095269522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,5120,0.01661777827474806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,512,5120,0.01237422227859497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,6144,0.022557333111763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,32,0.005570666657553778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,4096,0.05160177747408549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,512,0.015480889214409722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,3584,0.007972444097201029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,6144,0.0258542216486401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1024,1536,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,5120,0.010911111202504901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,2048,0.005234666582610872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,2048,0.01463644372092353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,12288,0.03870844509866502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,3072,0.02609777781698439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,4096,0.017815111411942374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,256,0.003167111012670729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,5120,0.02847733431392246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,5120,5120,0.035270220703548856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,768,0.01648000048266517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,16384,0.04258666767014397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,3584,0.007634667058785756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,3584,0.02253155575858222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,8192,0.010664888554149203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,256,0.0029528888149393927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,65536,1.565472920735677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,10240,0.05669777923160129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2560,10240,0.015524443652894763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,128,0.014845333165592618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,2048,0.018201778332392376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,65536,0.18223822116851807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,2560,0.06857866711086698
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,1536,0.00832088871134652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,5120,0.01720977822939555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,4096,0.009930666950013902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,768,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,32,0.005602666487296422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,32,0.003061333257291052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,16384,0.020615999897321064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,1024,0.008827555510732863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,256,0.012718222207493253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,2048,0.0052044445441828836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,768,0.010703999963071613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,7168,0.02715555495686001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,2560,0.02198133369286855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,2560,0.07029422124226888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,4096,0.01512977812025282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,12288,1024,0.009498666558000777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,16384,0.04506577716933357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,8192,0.010758221977286868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,2560,0.07852444383833143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,2560,0.021964444054497614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,8192,0.03621600071589152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,4096,0.009663999908500248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,6144,0.18769333097669813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,256,0.015293333265516492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,256,768,0.005687110953860813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,5120,0.08184533649020724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,6144,0.019926221834288705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,512,0.015697777271270752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,65536,0.07832177480061848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,256,0.017157332764731515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,768,0.015312888556056552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,16384,12288,0.04055644406212701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,5120,0.008341333104504479
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3584,1024,0.004222222086456087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,4096,0.01551822159025404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,65536,7168,0.16127555900149876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1536,7168,0.01033688916100396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,2560,0.011243555280897351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,1536,0.004579555657174853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,1024,0.00830666638082928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,512,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,2048,0.016198222835858662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,3072,0.007343110938866933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,10240,0.0069848886794514125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3584,8192,0.015299555328157214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,32,0.003810666501522064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,6144,65536,0.21013600296444365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,32,10240,0.0120328888297081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,8192,5120,0.39527199003431535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,2048,0.010823999842007955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,5120,0.008286221987671321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,10240,0.018623999423450895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,256,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,16384,0.1363324456744724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,1024,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,64,0.00351555562681622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,3072,0.007283555964628856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,6144,0.010321777727868822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,7168,0.029150221082899306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,10240,0.07852088742785983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,4096,0.04422933194372389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,65536,0.09253066778182983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,16384,0.0413315561082628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,4096,10240,0.04105777872933282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3072,1024,0.005935110979610019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,7168,3584,0.025624000363879736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2048,6144,0.020703110429975722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,1024,0.014873777826627096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,4096,0.018918222851223417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,2560,0.02606222199069129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,256,0.018246221873495314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,12288,0.3295519881778293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,7168,0.022984888818528917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,256,0.005625777774386936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,768,0.0035128887328836652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,512,0.015430221954981485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,12288,0.03714311122894287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,128,0.006261333409282897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,512,0.005411555369695027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,1024,0.008644444247086843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,5120,0.011376889215575324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,5120,0.019525332583321463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,12288,0.21403022607167563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,7168,0.018604444132910836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,128,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,8192,0.027492443720499676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,2560,0.019276445110638935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,3072,0.007551111280918121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,64,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,32,0.007582222421964009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,6144,0.009353777600659264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,3584,0.007314666277832455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,10240,7168,0.0314319994714525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,5120,0.00979288915793101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,65536,0.08148088720109728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,10240,0.08558311065038045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,6144,0.05111289024353027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,4096,0.01690933273898231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,4096,0.008565333154466417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,12288,0.013117333253224691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,128,0.005650666852792104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,512,0.005597333527273602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,64,5120,0.00866844422287411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,1024,0.013464888764752282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,2048,0.008976000050703684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,1024,0.016812443733215332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,64,0.0028764444092909494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,6144,0.023710222707854375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,32,1536,0.006232000059551663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,5120,0.08020977841483222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,512,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,3072,0.0158595558669832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,768,0.0035031110876136353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,768,0.00590844452381134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,8192,0.011029333704047732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,7168,128,0.005215999980767568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,10240,0.014752000570297241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,12288,0.02027466727627648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,128,0.0028488888508743713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,16384,0.03890044490496317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,5120,0.016383111476898193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,3584,0.007050666544172499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,2560,0.006160000132189856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,16384,0.3775324556562636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,1024,0.0073902226156658595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,2048,0.011338666909270816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,3072,0.012410666379663678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,65536,0.09380000167422825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,1536,0.017103999853134155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,2560,0.018745778335465323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,65536,32,0.016137777103318107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,256,0.011787555283970304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,3584,0.007684444387753804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,128,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1024,16384,0.02127200033929613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,3072,0.010007999837398529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,1536,0.008018666671382057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,3072,0.031680001152886286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,128,0.0058808889653947614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,16384,10240,0.05351110961702135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,2560,0.006216889040337668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,16384,0.07606310976876153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,16384,16384,0.0676835576693217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,256,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,768,0.004607111215591431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,65536,0.061928888161977134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,512,0.015106666419241162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,1024,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,65536,6144,0.057598219977484814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,6144,0.07224444548288982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,512,0.005950222412745158
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,12288,0.02143999934196472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,7168,0.04177688890033298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,256,0.01553066737122006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,1024,0.017874666386180453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,3072,0.014138667119873894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,1024,0.004550222307443619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2048,12288,0.028007109959920246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,5120,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,16384,0.014768888552983602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1536,3584,0.007272889216740926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,7168,0.0199626667631997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,5120,0.12219466103447808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,65536,0.09901422262191772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,8192,0.020883555213610332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,5120,0.00832622249921163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,1536,0.014748444159825643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1024,16384,0.01979377700222863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,8192,768,0.003924444317817688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,12288,0.024404444628291663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,12288,0.04390666551060147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,512,0.07058222426308526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,12288,0.06925688849555121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,3072,0.01365600029627482
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,10240,256,0.013883555928866068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,32,0.005246222019195557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,512,0.007276444799370236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,768,512,0.015126221709781222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,7168,0.029321776496039495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,12288,0.007000889215204451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,16384,0.012815111213260226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,6144,0.009675555759006077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,2048,0.006624889042642381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,16384,0.12134844726986355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,64,0.006751110984219446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,2560,0.047592888275782265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,2048,0.027635556128289964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,64,0.0027866665687825945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,512,0.010057777994208867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,3072,0.008213332957691616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1536,16384,0.03185866607560052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,16384,0.01684977776474423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,3072,0.014183110660976835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,64,0.02166755497455597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,512,0.0032026666320032547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,12288,0.040319111612108015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,32,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,65536,0.15531199508243138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,10240,0.018057778477668762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,2048,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,3584,0.01924800044960446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,32,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,768,0.029185778564876978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,3584,3072,0.028365333875020344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1536,3584,0.007656888829337225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,6144,0.010682666467295753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,64,0.0027786667148272195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,128,0.014858666393491956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,8192,0.015409777561823526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,10240,0.028207110034094915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,5120,0.0199608885579639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,128,0.0052586665583981415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,64,5120,0.008736888567606607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,256,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,1024,0.011593778100278644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,32,12288,0.02490577763981289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,12288,0.5490604506598579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,512,0.008616000413894653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,3584,0.021810667382346258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,128,0.009316444396972656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,16384,5120,0.04292622208595276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,65536,0.12931999895307753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,2560,0.022872888379626807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,10240,0.019175999694400363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,1024,0.018193778064515855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,10240,0.02611200014750163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,768,0.02071288890308804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,65536,0.18335999382866752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,12288,0.048284444544050426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,10240,0.03311911225318909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,10240,0.025352888637118872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,7168,0.0163137780295478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,5120,0.12309600247277154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,128,0.005527110977305307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,2560,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,768,0.0035048888789282907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,1536,0.006309333360857434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,6144,0.02017866737312741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,128,0.0177528891298506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,10240,0.016019556257459853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,256,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,3584,0.011223110887739392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3072,3072,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,8192,0.04020355476273431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,6144,0.026660444007979497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,3072,0.01888444523016612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,256,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,12288,10240,0.8936275906032987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,32,0.004620444443490771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,32,8192,0.022184888521830242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,65536,0.2355395687950982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,768,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,32,6144,0.009464888936943477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,12288,0.039326220750808716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,2560,0.015445333388116626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,1536,0.023344889283180237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,1536,0.006286222073766921
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,8192,0.10874311129252116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,1536,0.031324442889955305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,65536,12288,0.21083645025889078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,12288,0.01536266671286689
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,8192,0.045464889870749585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,65536,32,0.026384888423813715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,128,0.005558222118351195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,65536,0.3809013366699219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,1536,0.008583111067612966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,6144,0.009332444104883406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,4096,0.007987555530336168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,1024,0.006248000181383557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,10240,0.0958942241138882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,64,0.005649777750174205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,3584,0.014088888963063559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,12288,0.029164443413416546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,128,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,512,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,2560,0.009710222482681274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,3584,0.010148444109492833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,768,0.006191111273235745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,65536,0.2263368765513102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,32,0.010768888725174798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,2560,0.008427555362383524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,3072,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,2048,10240,0.1830728848775228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,1024,0.007344000041484833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,256,0.005245333330498801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,3072,0.013129777378506131
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,65536,0.1379688845740424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,3584,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,65536,0.055734223789638944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,5120,128,0.005961777849329843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,16384,0.00942488925324546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,10240,0.11037600040435791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,2560,0.012486221889654795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,12288,0.02732266651259528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,10240,0.007607999775144789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,16384,1.759476449754503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,65536,1.9723307291666667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,1024,0.019069333871205647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,5120,0.012426666915416718
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,10240,0.026326222552193537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,10240,0.10715999868181016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,1536,0.004559999952713649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,512,0.015447111593352424
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,5120,0.016529776983790927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,7168,0.009353777600659264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,512,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,1536,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,7168,0.00795377790927887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,128,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,3584,0.02328088879585266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,5120,7168,0.017083555459976196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,5120,0.006608888920810487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,256,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,32,0.00282577777074443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,3072,0.007274666594134436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,16384,128,0.0028719999310043124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,10240,0.015279110934999255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,3072,0.009342222577995723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2560,65536,0.0650577810075548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,10240,65536,0.14814666906992593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,2048,0.0170151111152437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,2560,0.03491910960939195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,512,7168,0.00962933318482505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,2560,32,0.003121777748068174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,8192,0.019295111298561096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,10240,0.018251554833518136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,8192,0.012791111237472959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,2560,0.015511999527613321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,65536,0.1439324484931098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,3584,0.2060257726245456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,256,0.00452622233165635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,5120,0.011477332976129321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,3072,0.010575110713640848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,3072,0.04870133267508613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,2048,0.015791111522250705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,3584,0.014807111687130399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,65536,0.02625688910484314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,4096,0.021563554803530376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,10240,0.025759999950726826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,1024,256,0.0059262220230367445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,1536,0.006937777830494775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,3072,0.0562737782796224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,64,8192,0.06552355819278292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,512,0.003546666767862108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2048,8192,0.02388266722361247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,16384,0.024323556158277724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,2048,0.023345778385798138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,768,2048,0.00656355544924736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,8192,0.026130666335423786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,2048,0.011776888536082374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,2560,0.0119982221060329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,3584,0.029222223493787978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,3072,0.007191110816266801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,5120,0.008967999782827165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,2560,0.01956977778010898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,8192,0.016093333562215168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,3584,0.008972444468074376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,512,0.0035048888789282907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,1024,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,768,0.0038408889538711975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,8192,0.031368000639809504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,256,768,0.006237333433495627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,7168,0.012677333421177335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,768,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,128,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,8192,0.02422133253680335
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,512,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,128,0.03677599959903293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,256,0.02404622236887614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,3072,0.011341333389282227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,128,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,4096,0.012054222325483957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,6144,0.114738663037618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,768,5120,0.009000000026490953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,16384,0.04874755607710945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,65536,3584,0.039228445953793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,6144,0.02918044394916958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,10240,0.01571911076704661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1536,65536,1.0838186475965712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2560,2560,0.011436444189813403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,512,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,16384,0.19562489456600615
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,6144,8192,0.01520088811715444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,6144,0.017474666237831116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,4096,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,128,0.004870222260554631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,8192,0.0875955555174086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,5120,0.04775289032194349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,7168,0.018000000052981906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,3584,0.00832622249921163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,16384,0.04318222072389391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,5120,0.02695111102528042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,256,0.015166223049163818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,6144,8192,0.017179555363125272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,2048,0.014130666851997375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,7168,1536,0.01757244434621599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,3072,0.009302222066455418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,2048,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,12288,0.014454222387737699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,7168,8192,0.023363555471102398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,1536,0.006579555571079254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1536,128,0.014760888285107084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,1536,0.004567111117972268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2048,1024,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,4096,0.018790221876568265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,512,0.005629333357016246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,64,32,0.0028648889727062653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,16384,0.015494222442309061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,256,0.019575110740131803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,3584,0.013568888935777875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,1536,0.006167111297448476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3072,7168,0.02236533330546485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,512,4096,0.007680888805124495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,12288,0.013447110851605734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,65536,0.4111662175920274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,12288,0.015784000356992085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,768,0.026399999856948853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,256,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,3584,0.011048888994587792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,64,0.005930666708283954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,2560,0.009644444617960189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,256,0.01480888823668162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,16384,0.7790791193644205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,256,0.0053075556125905775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,10240,0.03750666644838121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,3072,0.00960888879166709
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,2560,0.00723555518521203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,65536,0.06003733476003011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,768,0.08633333444595337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,3584,0.09238400061925252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,32,1024,0.0038773334688610504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,16384,0.02104622291194068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,768,0.0035617777870761026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,128,0.003023111157947116
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,16384,0.015499555402331881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,10240,0.024547555380397375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,7168,0.03815999958250258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,3584,0.032300445768568255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,512,0.0373306671778361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,128,0.003232888877391815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,512,0.007303111255168915
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,10240,0.016452444924248587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,2048,0.009953777823183272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,768,0.04778666628731621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,128,0.022006221943431433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3584,1024,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,256,0.01757244434621599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,512,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,128,5120,0.008983111215962304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,3584,0.007322666545708974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,5120,0.013289777768982781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,64,0.00601333338353369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,2560,0.019900444481107924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,1536,0.17956177393595377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,12288,0.014025777578353882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,768,0.014787554740905762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,12288,0.02405599918630388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,128,0.005233777893914117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,256,0.016536000702116225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,1024,0.007300444775157505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,128,0.02239822182390425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,1536,0.013764444324705334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,3584,0.034633778863483004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,12288,0.3197724554273817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,128,0.006575111299753189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,12288,0.05036266644795736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,6144,0.015622221761279635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,2048,0.014172444740931192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,4096,0.014133334159851074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,16384,0.35414934158325195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,128,0.003922666526503033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,16384,0.01666844387849172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,5120,0.020640888147883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,5120,0.02404533326625824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,7168,1536,0.007429333196745978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,6144,0.03436355458365546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,64,0.004586666822433472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,512,0.015103111664454142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,32,0.012038222617573209
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,32,0.007183999651008182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,2048,0.007923555870850882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3072,5120,0.00831200016869439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,8192,128,0.038839111725489296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,128,0.03367999858326382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,1024,0.0041928887367248535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,2560,0.006310222049554189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,6144,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,3584,0.029655112160576716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,2560,0.05855111281077067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,10240,0.017096888687875535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,2560,0.010423999693658618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,1024,0.07977510823143853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,7168,2560,0.0073315559162033936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,4096,0.008347555167145198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,2048,0.007015110717879401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,7168,0.015523556205961438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,10240,0.01349422252840466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,4096,0.18415466944376627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,3584,0.041658666398790144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,64,0.01554666625128852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2560,16384,0.021661332911915247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,64,0.005663110978073544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,7168,0.00795111142926746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,3072,0.00738666703303655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,1536,0.0069120000633928515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,7168,0.016202666693263583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,6144,1536,0.04912177721659342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,2560,0.013440888788965015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,1536,0.009031111167536842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,1536,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,32,0.0028231110837724474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,3584,0.01867377758026123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,768,0.014440889159838358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,2560,0.007215110792054071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,8192,0.024159111910396155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2560,256,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,10240,0.0853688915570577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,2560,0.01481244464715322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,8192,0.01963733302222358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,65536,0.10983910825517441
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,65536,0.04584177666240268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,8192,0.025399999486075506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,64,0.002899555489420891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,256,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,65536,10240,0.10063111119800144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,128,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,1024,0.007631111476156447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,1024,0.017910222212473553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3072,16384,0.021479111578729417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,5120,0.02664711078008016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,16384,2048,0.013442666994200813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,12288,0.06590844525231256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,8192,0.03298133280542161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,768,0.004256000121434529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,1024,0.011707555916574268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,12288,0.019925332731670804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,64,0.00591555568906996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,65536,0.06921422481536865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,768,32,0.006671999891599019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2560,768,0.007245333658324347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,256,1024,0.012392000191741519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,6144,10240,0.019328888919618394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,16384,0.03357155455483331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,256,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,512,0.00980177770058314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,4096,0.019474666979577806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,3072,0.03212711215019226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,2048,0.008608000146018134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,32,0.0031786666562159858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,512,0.0069120000633928515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,8192,0.009185777770148383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,512,0.01204711116022534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,64,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,65536,0.07574933105044894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,768,0.017455110947291057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,12288,0.11039288838704427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,256,0.006802666518423293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,6144,0.012433778080675336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,5120,0.009240888886981541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,8192,0.03236088818973965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,4096,0.011403555671374003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,2560,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,10240,0.012234666281276278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,4096,0.2036177714665731
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,4096,0.016563556260532804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,128,0.007321777443091075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,64,0.003206222214632564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,16384,0.018583110637134977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,512,1024,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,128,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2048,65536,0.0634017785390218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,4096,16384,0.02255288925435808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,12288,0.013581333061059317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3584,256,0.023353776997990076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,1536,0.008730666504965888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,8192,0.06219289037916395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,65536,0.12002044253879124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,512,0.005906666732496685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,32,0.0028924445311228433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,3072,0.007629333270920648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,10240,7168,0.06346399916542901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,10240,0.458346684773763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,128,16384,0.015467555986510383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,3584,0.010712888505723743
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,512,0.007632888853549957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,256,0.00759733302725686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,128,0.0035048888789282907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,65536,0.06701955530378553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,5120,65536,0.07309866613811918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,7168,0.020911999874644812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,768,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,4096,0.04082311193148295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,128,0.005655111124118169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,6144,0.014493332968817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,6144,0.04596266812748379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,3584,0.014610666367742749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3072,7168,0.012466666599114736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,7168,0.010904889139864178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,64,0.008294222255547842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,12288,0.013457777599493662
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,6144,0.013553777502642738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,1536,0.01551555593808492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,16384,0.024254222710927326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,4096,0.018554665976100497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,1536,0.06956888569725884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,1024,0.0059057776298787855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,1024,0.015180443723996481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,16384,0.02068533334467146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,5120,0.010026666853162978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3584,256,0.003544888976547453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,7168,0.009657777845859528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,8192,0.044885334041383534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,512,0.018944889307022095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,128,0.0052577778697013855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,7168,2048,0.028471999698215063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,65536,0.0917466680208842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,5120,0.05792444282107883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,768,0.00590133335855272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,768,5120,0.010647999743620554
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,8192,0.011400000088744693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,256,0.009362666971153682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,16384,10240,0.20643644862704805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,7168,0.02310577697224087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,2560,0.007997333175606197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,7168,0.02894222074084812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,65536,0.1541715595457289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,256,0.003176888864901331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,16384,128,0.0031191110610961914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,10240,0.012081777883900536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,65536,0.056574225425720215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,10240,256,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,12288,0.08007111152013142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,3072,0.011309333145618439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,16384,0.05595911211437649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,1024,0.006653333289755716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,65536,0.06389777527915107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,16384,0.007987555530336168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,3584,0.015186667442321777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,768,0.009018667042255402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,4096,0.008000888758235509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,10240,0.03042311138576931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,5120,0.05797600083880954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,5120,0.016534222496880423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,16384,0.023423110445340473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,5120,0.02747822304566701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,768,0.014793778459231058
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,12288,0.06057688925001356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,65536,3.7115118238661022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,4096,0.00868622213602066
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1536,2560,0.01513777838812934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,512,0.00452977791428566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,256,0.0035235554807715942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,64,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,2560,0.009664889011118147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,2048,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,5120,0.01092266705301073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,2048,0.017346666918860543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,65536,0.06496089034610324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,512,3584,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,64,0.0032115555885765287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,2560,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,1536,0.006146666490369373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,6144,0.011696000066068439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,65536,0.022707555029127333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,32,6144,0.005948444621430502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,4096,0.01616533266173469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,10240,0.011772444678677453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,64,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,3072,0.010008888940016428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,4096,0.010030222435792288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,4096,0.015407111909654407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,128,0.005600000007285013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,512,0.005586666779385672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,10240,10240,0.4777955479092068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,7168,0.021418665846188862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,16384,12288,0.04083999991416931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,65536,0.2364177703857422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,65536,0.06641511122385661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,512,0.004369777937730153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,2048,0.012836444709036084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,10240,0.03031022349993388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,5120,0.02644177774588267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,32,128,0.008042666647169325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,6144,0.017177777157889474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,12288,0.022786666949590046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,768,0.016865778300497267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,8192,128,0.015063999427689446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,128,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,65536,7168,0.09863377941979302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,6144,0.020459555917316012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,5120,0.12831021679772273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,5120,2048,0.05423288875155979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,1536,0.026816000541051228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,2048,0.008989333278603023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,2560,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,8192,8192,0.033273776372273765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,65536,0.14616000652313232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,10240,0.011711999773979187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,128,0.003154666680428717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,768,0.004219555606444676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,3584,0.008286221987671321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,2048,0.014820444915029736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,8192,0.08586400085025364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,7168,0.10857422484291925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,2560,0.006976000136799282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,65536,0.015000889698664347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,128,3072,0.007228444847795699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,6144,0.018592000007629395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,64,0.003535110917356279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,2560,0.020434666011068556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,512,0.008966222405433655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,1024,0.004571555389298333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,4096,0.03312799996799893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,12288,0.02237155536810557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,5120,0.00905600024594201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,4096,0.017840888765123155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,128,0.005613333235184352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,768,0.003974222060706881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,16384,0.054678221543629967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,2560,0.006894222150246303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,128,0.0028862222615215513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,1536,0.006351999938488007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,512,0.02641155487961239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,3584,0.03900266687075297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,768,0.006224888894293044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,3584,0.5350649091932509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,32,0.006224000205596288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2560,7168,0.021410667233996924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,10240,0.012131555212868584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,8192,0.06358310911390516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,1536,0.0047600000268883174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,7168,0.02994933393266466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,12288,0.18942488564385307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,65536,0.14345955848693848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,2048,0.017329777280489605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,3584,0.01422133379512363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,4096,0.010388444695207808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,1024,0.006022222340106964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,3072,0.012096888489193387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,7168,0.054437332683139376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,7168,0.006262222097979651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,12288,0.01921599937809838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,32,6144,0.00941066692272822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,128,0.006933333145247565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,128,0.0032133333798911837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1024,1536,0.004635555462704765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,3584,0.016607999801635742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,1024,0.013447999954223633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,8192,0.010703999963071613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,768,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,1024,0.004171555654870139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,128,0.003126222226354811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,4096,0.009072889056470659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,65536,0.061848892105950244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,3072,0.022691556149058874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,16384,0.00867911097076204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,512,0.004572444491916233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,1024,0.005945777727497949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,3072,0.011429333024554782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,6144,0.030751112434599135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1536,768,0.00489688871635331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,3072,0.011644444531864591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,2048,0.0074728892909155945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,1536,65536,0.5848151312934028
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1536,4096,0.02126666737927331
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,4096,0.010408889088365767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,6144,0.1290231148401896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,6144,0.01756800048881107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,1024,0.03868710994720459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,1024,0.026199110680156287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,10240,0.305967993206448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,6144,128,0.0069360000391801195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2560,8192,0.01071111112833023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,2560,0.007317333585686154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,5120,0.021879111727078755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3584,64,0.013859555953078799
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,128,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,5120,0.16085422039031982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,12288,0.17715377277798125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,7168,0.6400702264573839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,3584,0.01312355531586541
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,7168,0.03865422142876519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,8192,0.03694044550259908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,64,0.03509155578083462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,32,0.006978666616810693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,7168,0.03766577773623996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,2560,0.008640888664457533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,2048,0.0070293330483966405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,4096,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,128,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3072,7168,0.024487111303541396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,65536,0.06412355767356025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,10240,0.03495288888613383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,65536,128,0.011707555916574268
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,7168,0.013829333086808523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,3072,8192,0.029078221983379785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,7168,768,0.00795999997191959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,16384,0.03660888804329766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,7168,0.00944622192117903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,3072,0.01717866626050737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,10240,0.04219288958443535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,8192,0.01498044364982181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,7168,0.030441777573691473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,7168,0.16533154911465114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,16384,0.021749332547187805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,8192,0.01648177703221639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,8192,10240,0.053915556934144765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,1536,0.005953777581453323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,1536,0.006245333287451003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,512,0.031390223238203264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,10240,0.015788444214397006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,65536,0.2540382279290093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,128,0.00627377794848548
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,1024,0.01648977730009291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,4096,0.058789332707722984
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,1536,0.00630666646692488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1024,2048,0.006994666324721442
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,65536,0.0649626652399699
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,16384,0.02304800020323859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3584,256,0.0028604444944196274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,5120,0.01755555636352963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,32,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,6144,0.08371555805206299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,65536,0.6792435646057129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,12288,0.02275466587808397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,768,768,0.021254221598307293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,65536,0.35287200080023867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,8192,0.016775111357371014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,6144,0.0136817776494556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,5120,0.008633777499198914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,65536,0.13116799460517034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,64,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,2560,0.02433066732353634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,2560,0.026100445124838088
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,1536,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,16384,0.026625777284304302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3584,10240,0.01512888901763492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,1536,0.006066666709052191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,4096,0.04996888836224874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,12288,0.02752799954679277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,6144,0.034274667501449585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,8192,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,16384,0.016738666428460013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,3072,0.007333333293596904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,65536,8192,0.16752000649770102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,7168,0.02257333364751604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,128,0.014789332946141561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,128,0.005588444570700328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,65536,0.15355200237698025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,5120,2560,0.00867911097076204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,5120,0.012433778080675336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,16384,0.021634666456116572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,10240,0.028128001425001357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,6144,0.01772622267405192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2048,65536,0.06301510996288724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,768,0.013615111509958902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,128,0.0031884445084465873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,1536,0.03320533368322585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,256,0.02238399949338701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,3072,0.025832888152864244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2560,512,0.01404800017674764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,2560,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,2048,0.009944000177913243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,7168,0.009321778184837764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,8192,3072,0.011329777538776398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,4096,12288,0.057004445128970675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,768,0.01888799998495314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,5120,4096,0.019174221489164565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,768,5120,0.008952889177534316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,7168,0.012759110993809171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,8192,0.01512888901763492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,1536,0.006305777778228124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,10240,0.020761777957280476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,256,0.006268444574541516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,16384,0.07874577575259738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,12288,0.04155733187993368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3584,3584,0.011021333436171213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,7168,1024,0.00794577764140235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,1536,0.0069617778062820435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,3584,0.039134220944510564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,2560,0.007292444507280986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,1536,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,512,512,0.006263111200597551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,10240,512,0.018199111024538677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,6144,0.017720000611411203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,512,0.016159999701711867
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,10240,0.03765866822666592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,1024,0.06856889194912381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,3072,0.024751999311976965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,3584,0.05437066819932726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,2048,0.006983111302057902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,1024,0.011742221812407175
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,768,0.005807111246718301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,12288,0.013455111119482251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,1024,0.007364444434642792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,65536,0.18798577785491943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,3072,0.006918222126033571
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,10240,0.02677511175473531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,256,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,768,0.00555377784702513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,16384,1536,0.03427911135885451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,10240,5120,0.02219999995496538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,2560,0.007613333563009898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,2560,0.007323555648326874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,2048,0.010059555371602377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,512,0.04959288901752896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,768,1024,0.008016889293988546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,64,8192,0.03978133201599121
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,10240,0.01237422227859497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,768,2560,0.020516445239384968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,8192,0.024319110645188227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,5120,0.00869333330127928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,8192,0.011408888631396823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,768,0.005882666756709416
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,12288,0.030346665117475722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,5120,3584,0.009580444130632613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,1536,0.010663111176755695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,1024,0.006213333457708359
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,7168,0.010623111493057676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,128,0.00796977761718962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,6144,0.017657778329319425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,8192,0.014351111319330005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,4096,0.02329866588115692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,2048,0.006220444622966979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,6144,12288,0.019718221492237516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,10240,128,0.006173333360089197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,12288,0.02343822187847561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,16384,768,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,512,0.04302488764127096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,6144,0.011401777466138205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,5120,0.00833333366447025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,4096,0.020011555817392137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,3584,0.013803555733627744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,65536,0.09701866573757595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,64,32,0.005564444594913059
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,4096,12288,0.019262222780121695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,3584,0.008633777499198914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,10240,0.06293155749638875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2560,6144,0.008976000050703684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,16384,0.012726222475369772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,256,0.01907911068863339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,768,0.017190222938855488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,2560,0.008917333351241218
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,2048,0.006528000036875407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,65536,0.15705245070987278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,3584,0.011738667057620155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,4096,0.025103999508751765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,1024,0.013422222601042854
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,512,0.01275466630856196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,5120,0.03269688950644599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,2560,0.012624888784355588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,1024,0.007987555530336168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,10240,0.035754667388068304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,7168,0.015142222245534262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,8192,7168,0.013738666971524557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,65536,0.807221359676785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,128,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,1536,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,1536,0.006278222219811545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,512,0.04363377888997396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,5120,0.05760266383488973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,6144,0.009290667043791877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,2048,0.00544888898730278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,1024,0.003988444391224119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,6144,0.009188444250159794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,6144,10240,0.01869066721863217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,1536,0.008016889293988546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,5120,0.010317333042621613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,64,0.004000888930426704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,512,0.014482667048772177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,1536,0.02715733316209581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,2560,0.026143999563323125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,128,0.012044444680213928
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,1024,0.00562755556570159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,2560,0.012006222373909421
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,16384,0.06596444712744819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,12288,0.03379111157523261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,4096,0.022871999277008906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,8192,0.014484445254007975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,6144,0.009322666459613377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,3584,0.0075937774446275495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,12288,0.024694222542974684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,6144,0.0664915574921502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,10240,1024,0.01926933394538032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,768,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,8192,0.010779555473062726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,6144,0.009374221993817223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,64,3584,0.007607111500369177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,2048,0.01293244461218516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,10240,0.016480889585283067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,768,8192,0.01110222190618515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,5120,8192,0.0162124451663759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1024,128,0.0032053333189752367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,1536,0.012453333371215396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,7168,0.18050400416056314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,2048,0.021465778350830078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,64,0.0029022221763928733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,128,0.0052328887912962176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,6144,0.07560266388787164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,1536,0.00488800017370118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,1024,0.013908444179428948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,3072,0.020974222156736586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,2048,0.005214222189452913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,256,0.005572444448868434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,32,2048,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,768,0.018212444252438016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,10240,65536,0.14083200030856663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,3584,0.009480888644854227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,6144,1024,0.006118222243256039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,2048,0.015416888727082146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,16384,0.04303466611438327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,3584,0.012448888685968188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,10240,1.6047013600667317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,1024,0.012173333101802401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,768,0.015174221661355762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,12288,0.013482666677898832
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,8192,0.18965510527292886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,3072,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1536,256,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,6144,0.010379555324713389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,6144,0.011323555476135679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,16384,0.1922924386130439
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,2560,0.13867555724249944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,4096,0.008935111264387766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,5120,0.03703022334310744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,3584,0.01313066648112403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,2560,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,512,0.0041653331783082755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,4096,0.007598222129874759
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,1024,0.01312800000111262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,2048,0.01685333251953125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,1024,0.003847111016511917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,4096,0.08798311154047649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,3072,0.013471999930010902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,8192,0.013460444079505073
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,768,0.01628177695804172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,7168,0.025394666526052687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,512,0.12218133608500163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,8192,0.030127998855378892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,7168,0.024002666274706524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,8192,0.03670755690998501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,5120,0.03500266538725959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,65536,1536,0.06937599844402738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,64,5120,0.005912000106440649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,512,0.010462221999963125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,1536,0.009687110781669617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,7168,0.010686222049925061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,1536,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,7168,16384,0.03958755400445726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,1024,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,8192,0.025171554750866357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,3072,0.005962666538026597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,128,0.004872000051869286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,4096,0.04517066809866163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2048,2560,0.015152888165579902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,3072,0.015041778484980265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,16384,0.007438222567240397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,64,65536,0.10578133662541707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,32,0.008003555238246918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,2048,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,512,1536,0.006081777728266186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,4096,0.03036622206370036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,768,7168,0.018562666243977018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,1536,0.030470222234725952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,6144,64,0.003933333274390963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,2560,0.006055999961164262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,16384,0.045238223340776235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,10240,0.011698666546079846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,128,0.005264889034960005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3584,64,0.011399110986126794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,128,0.006638222270541721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,6144,0.027789334456125896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,3072,0.012993777791659037
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,2048,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,512,0.0038817777401871155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,512,0.005604444278611078
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3584,2048,0.007089777953094906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,128,0.004832889056868023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,2560,0.007263110743628607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,6144,0.010919111470381418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,12288,3584,0.015985776980717976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,512,0.0035573335157500375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,768,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,64,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,4096,768,0.005923555543025334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,4096,0.011746666497654386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,128,0.00291377781993813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,64,0.006315555423498154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,1536,0.006622222148709827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,12288,0.04010666741265191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,6144,0.04537422127193875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,2048,0.006610666712125142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,6144,0.016536000702116225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,65536,0.9161351521809896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,6144,0.05383377936151293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,5120,0.0069431112044387394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,128,0.0059368887709246735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,6144,0.01789866718981001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,65536,0.09113511112001206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,2560,0.006976889239417181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,3072,0.012366222010718452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,4096,0.011373333632946014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,10240,0.015311110350820752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,16384,0.04610310991605123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,7168,0.01853511068556044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,256,0.00589155571328269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,10240,0.03336000111367967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,4096,0.007663110891977946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,512,0.018631999691327412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,4096,0.006272000157170826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,7168,0.18482222821977404
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1536,1536,0.014807111687130399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,32,0.004523555437723796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,2560,0.010087110930018956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,6144,0.013667555318938361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,7168,0.0103502223889033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,32,0.005919111271699269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,5120,0.009010666774378883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,7168,0.010066666536860997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,10240,0.015800888339678448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,32,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2048,3072,0.01648177703221639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,768,0.005678222411208683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,12288,0.019117333822780185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,7168,2560,0.07724977864159478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,32,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,10240,0.0397093329164717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,3072,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,2560,0.013454222016864352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,8192,0.01475288967291514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,16384,4096,0.016184000505341422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,256,16384,0.016217778126398723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,6144,0.01056977775361803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,3072,0.016507556041081745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,12288,0.018255111243989732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,2560,0.028024001253975764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,256,0.006617777877383762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,2560,0.007640889121426477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,16384,0.15425954924689397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,128,512,0.01274400038851632
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,2560,0.007636444436179266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,10240,0.016248888439602323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,5120,512,0.04287822379006279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,10240,0.01906222270594703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,16384,0.2613831096225315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,32,0.00286311118139161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,2560,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2048,32,0.009726222190592024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,3072,0.0069271110826068455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,256,0.005722666780153911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,2048,0.013790222505728403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,3072,0.0086986662613021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,512,0.013228444589508904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,65536,2560,0.13087199793921575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,6144,0.009367999931176504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,256,0.0052684446175893145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,10240,0.015812445018026564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,1536,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,768,0.006222222414281633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3584,256,0.007675555845101674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,8192,2560,0.0075697774688402815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,12288,0.007285333342022366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,512,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,10240,0.01143555591503779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,3072,0.007364444434642792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,2048,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,32,0.005985777825117111
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,5120,0.008417777717113495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,1024,0.01793066660563151
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,2560,0.007635555333561367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,1536,0.01956355571746826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,4096,0.008329778081840938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,2048,0.021938666701316833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,65536,0.02288977801799774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,256,768,0.005619555711746216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,256,0.0038862224254343244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,7168,0.017866666118303936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,256,12288,0.013404444687896304
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,10240,0.01661422186427646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,4096,0.023043556345833674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,6144,16384,0.062386665079328746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,7168,0.006587555425034628
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,1536,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,3072,0.008769777913888296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,512,65536,0.06211466921700371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,1024,0.007308444215191736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,3072,128,0.005949333310127258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,256,0.033360888560613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,10240,0.012113778127564324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,2560,0.008325333396593729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,8192,128,0.014113777213626437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,2048,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,6144,0.009024000002278222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,12288,512,0.009222221871217092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,16384,0.016548444827397663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,512,0.005946666830115848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,64,2048,0.006578666468461354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,512,0.0059279998143513995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,512,0.006598222172922558
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,8192,0.023455111516846552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,2048,0.018607111440764535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,10240,0.02020088831583659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,12288,256,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,2560,0.00793688909875022
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,65536,0.29304178555806476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,2048,0.030508445368872747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,3072,0.01016622202263938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,10240,0.015628443823920358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,7168,0.03153955605294969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,5120,0.013047110703256396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,768,0.004193777839342753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,8192,0.025313778056038752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,7168,256,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,5120,0.011289777855078379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,16384,0.20793422063191733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,8192,0.02165599995189243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,768,0.0038453332251972626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1024,16384,0.018538667096032035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,5120,0.020847110284699336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,16384,0.30513244205051
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,6144,0.009964444571071202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,65536,0.04796888762050205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,32,0.010598222414652506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,512,0.015809777710172866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,16384,3072,0.01716444392999013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,3072,0.030730668041441176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,256,0.0059573331640826324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,2560,0.007368889119890001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1536,3584,0.007521777517265744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,3584,0.027535999814669292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,7168,0.024712888730896845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,16384,2560,0.03208533260557387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,7168,0.013399110900031196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,512,0.014478221535682678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,768,0.00481244424978892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,128,0.00556355549229516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,2048,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,8192,2048,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,128,0.014455111490355598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,7168,0.014345778359307183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,768,0.022579555710156757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,4096,1024,0.014741332994567024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3072,6144,0.012748444245921241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,1024,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,12288,0.01648266613483429
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,128,16384,0.015832889411184523
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1024,2048,0.005609777652555042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,7168,0.02165066699186961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1536,12288,0.013416888813177744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,3072,0.012228444218635559
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,1536,0.020054222808943856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,7168,0.010315555665228102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,10240,0.1561111079321967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,2560,0.007311111523045435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,1536,0.01686311099264357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,5120,0.01752355529202355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,32,3072,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,10240,0.017297777864668105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,2048,0.006923555499977536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,3072,0.00776444458299213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,12288,0.032602667808532715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,128,2560,0.02276800076166789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,16384,65536,4.4496481153700085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,8192,0.03260533346070184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,128,0.008357333640257517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,128,0.00573333352804184
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,6144,0.021487111846605938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,128,65536,0.0528417792585161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,4096,5120,0.024903999434577093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1024,128,0.0247013337082333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,3072,0.03134222163094415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,2560,0.023048889305856492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,16384,0.03456711106830173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,16384,0.2078782187567817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,1024,0.003972444269392225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,65536,0.04494488901562161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,512,0.015589333242840238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,768,3072,0.008032889001899296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,4096,65536,0.06695466571384005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,7168,0.016608888904253643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,16384,3584,0.032383110788133405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,5120,0.18726578023698595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,768,0.01798311041461097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,768,7168,0.01239466667175293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,768,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,768,0.18390044901106092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,128,0.003277333246337043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,128,32,0.005712000032265981
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,16384,1024,0.024004444479942322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,2560,0.00927111092540953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,6144,0.0177875558535258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,16384,0.024516445067193773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,12288,0.0756853355301751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,64,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,128,0.007149332927332983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3584,3072,0.05797244442833794
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,768,0.005241777747869492
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,4096,0.01959733333852556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,65536,12288,0.23805422253078887
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,6144,0.013616888887352414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2560,7168,0.026102221674389307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3072,256,0.00591111100382275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,2048,0.04120444589191013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,6144,0.062255111005571157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,6144,5120,0.0155102229780621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,12288,0.022035555707083807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,4096,12288,0.019688000281651814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,4096,0.007279111279381647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,1536,0.004516444272465176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,3072,0.01756177842617035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,10240,0.015110222829712762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1536,7168,0.01923466722170512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,3584,0.036551998721228704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,2048,0.006591111007663939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,3584,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,7168,0.022308443983395893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,6144,0.011778666741318174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,2048,12288,0.02843377656406826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,65536,0.07585510942671034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,256,5120,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,7168,0.014136888914638095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,128,0.002875555513633622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,1024,0.011723555624485016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,5120,0.01239111108912362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,768,0.2178551091088189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,16384,2560,0.017641777793566387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,1536,0.014167111780908374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,512,0.0127368892232577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,8192,0.050824890534083046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,12288,0.021016000045670405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,12288,0.027443556321991816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,768,0.017521777086787753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,10240,3072,0.0359208881855011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2048,5120,0.018264888061417472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,32,0.0035208890007601846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,7168,0.01610844499535031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2560,2048,0.0052373334765434265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,2048,0.025891555680169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,8192,0.03926488757133484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,3072,0.024320888850424025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,2048,0.006623111251327727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,65536,768,0.01721244388156467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,3584,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,7168,0.009698666632175446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,128,0.003141333245568805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,16384,0.03358755509058634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,256,0.007620444728268518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,12288,0.038605332374572754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,768,0.007244444555706448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,4096,3072,0.007256000406212277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,10240,2048,0.022543110781245764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,7168,0.01407911048995124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,128,5120,0.008664000365469191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,7168,0.012106666962305704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,12288,0.24028176731533477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,8192,768,0.011392888923486074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,8192,2560,0.046239111158582896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3584,128,0.0151573336786694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,512,0.005202666752868229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,256,0.01069866700304879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,5120,0.01625066664483812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,2048,0.047616001632478505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3584,2048,0.020248000820477802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,1024,0.005798222290145026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,1536,0.015159999330838522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,65536,5120,0.1656551096174452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,256,0.003192000091075897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,512,0.0035626664757728577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,12288,32,0.0035422220826148987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,1536,0.006912888752089606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,3072,0.01458666721979777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,128,0.005559999909665849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,10240,0.02218755582968394
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,65536,0.08366044362386067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,768,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3072,768,0.016544888416926067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,32,0.002817777709828483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1024,5120,0.008573333422342936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,10240,0.011727111207114326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,6144,5120,0.01812977757718828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,768,0.016530666086408827
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,6144,0.07545689079496595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,512,7168,0.009696000152164036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,256,0.014846222268210517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,12288,0.046965334150526256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,2560,0.009008000294367472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,1536,0.007616000043021307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,6144,0.053357332944869995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,2560,1536,0.044847998354170054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1536,256,0.015462223026487561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,16384,0.008007999923494127
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,512,128,0.004576888763242298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,128,0.015120888749758402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,3584,0.007304000357786815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,10240,0.031255112753974065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,6144,0.010399111443095738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,128,0.0032142222755485107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,3072,0.017421333326233756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,8192,0.009145778086450364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,512,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3584,256,0.0031537777847713898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,10240,0.04127733243836297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,768,0.0069644442862934535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,256,256,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,1024,0.006262222097979651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,3584,0.028396444188223943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,12288,0.06439733505249023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,3584,0.0941457748413086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,128,0.015115555789735583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,128,6144,0.01348088930050532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,2560,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,1024,0.0041644444896115195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,2048,0.005717333406209946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,512,16384,0.028799109988742407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,6144,0.0399582220448388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2048,3584,0.0092604441775216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,2560,0.011982222398122152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,1536,0.007653333246707916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,10240,0.21423644489712187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,3584,0.1655528942743937
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,1536,0.012127111355463663
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,6144,3072,0.007286222444640265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,512,5120,0.01034400032626258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,6144,2048,0.017852443787786696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,2560,0.007063111497296228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,3584,0.025391111771265667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,12288,0.016922665966881644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,768,0.010751999914646149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,4096,0.009194666312800514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,2560,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,12288,128,0.0029093333416514923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,10240,8192,0.08628000153435601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,6144,5120,0.02288088864750332
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,3584,0.018058665924602084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,2048,0.01402311192618476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,12288,1024,0.01927911076280806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,7168,0.012140444583363004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3072,128,0.017842666970358956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,5120,0.07101688782374065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,1024,0.03605066736539205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,1536,0.004978666702906291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,16384,0.01939733326435089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,65536,0.13988355795542398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,7168,0.019891555110613506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,128,0.016491555505328707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,3072,0.007522666619883643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1024,2560,0.010675555302037133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,65536,512,0.05784622165891859
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,12288,0.047705776161617704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,7168,256,0.003907555507289039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,32,8192,0.01071111112833023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,8192,0.005271111097600725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,7168,0.0058871110280354815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,2048,0.014328888720936246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,16384,10240,0.06732088989681669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,64,0.0031724443866146933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,2048,0.005917333480384614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,256,0.01548266741964552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,5120,0.030663109487957422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,4096,0.05243200063705444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,10240,0.03895111216439141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,4096,0.02517333295610216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2048,8192,0.030568000343110826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,5120,0.03543466660711501
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,7168,0.10977866914537217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,3072,0.007605333295133378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,7168,0.009368889033794403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,5120,768,0.017490666773584154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,16384,0.01613066593805949
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,4096,6144,0.026527111728986103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,5120,0.00868266655339135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,12288,0.026064000195927087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1536,1024,0.01387466655837165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,3584,0.007325333025720384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,1024,0.009302222066455418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,32,0.012097777591811286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,7168,16384,0.04394577940305074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,5120,0.024115555816226538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,7168,32,0.005616888817813661
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,1024,0.014814222852389017
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,1536,8192,0.05586577786339653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,8192,0.015087111128701104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,65536,0.12633244196573892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2560,10240,0.03240799903869629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,3584,0.015895111693276297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,32,0.003165333428316646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,5120,256,0.0038986665507157645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2560,3584,0.00797333319981893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,512,0.020505777663654752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,768,0.006923555499977536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,12288,4096,0.03192889028125339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,65536,0.03322488731808133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,128,0.00795555528667238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,3584,0.011443555355072021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,5120,0.062154663933648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,768,0.024580443898836773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,4096,0.010370666782061258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,12288,0.02550933261712392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,256,0.006532444308201472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,5120,0.032807999187045626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,7168,0.0432168874475691
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,6144,256,0.014871999621391296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,32,1024,0.0038364442686239877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,32,0.00286666676402092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,2560,2048,0.01482311056719886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,128,6144,0.008998221821255153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,768,0.004552000098758274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,7168,12288,0.01884622209601932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,12288,0.018546667363908555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,128,32,0.0029226665695508323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,2560,0.0273635552989112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,12288,12288,0.05439466569158766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,6144,0.014875556031862894
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3072,1536,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,12288,5120,0.040087110466427274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,2560,0.0069297779765393995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,64,0.0032044444233179092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,5120,0.012785777449607849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,1024,0.006257777826653586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,65536,0.060106668207380504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,2048,0.03335288829273648
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,5120,0.02624977793958452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,8192,0.0107342220014996
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,12288,0.016884444488419425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,768,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,8192,0.03391733434465196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,5120,0.06490133206049602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,5120,0.011415111521879831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,768,2048,0.010509333676762052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,7168,0.0458586679564582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,65536,0.24951288435194227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,8192,0.026184888349639043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,4096,0.007655111451943715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,128,0.014828445182906257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,65536,0.024849777420361836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,1536,0.004994666824738185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,128,0.007328888608349695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,512,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,4096,0.010379555324713389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,6144,1536,0.010111110905806223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,3584,0.0076453329788313965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,512,0.005391110976537068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,65536,0.07223555776807997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,12288,0.03032355507214864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1024,8192,0.010690666735172272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,2560,0.014109333356221518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2048,12288,0.015528000063366361
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,8192,3584,0.025054223007626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,128,0.016311999824311998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,5120,0.02023555503951179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,1536,0.04979022343953451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,3072,0.05061866508589851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,7168,0.027246221899986267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,65536,0.5031626489427354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,12288,0.0338986681567298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,512,0.004882666799757215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,16384,0.05221688747406006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,5120,0.024679111109839544
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,128,0.0031306667046414483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,4096,0.015154666370815702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,12288,0.015800888339678448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,3072,0.028565333949195013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,2048,0.014966222974989148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,256,0.006996444529957241
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,4096,0.009743111001120673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,32,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,8192,6144,0.02734666731622484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,1536,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,128,10240,0.011718221836619906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,6144,0.013467555244763693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,5120,256,0.005617777920431561
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,6144,0.09802044762505426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,12288,256,0.012014221813943652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,64,0.0028373334142896864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,256,0.0031973332580592898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,5120,0.013087111214796701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,2048,5120,0.008643555144468943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,6144,0.027912888262006972
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,128,0.0052284445199701525
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,1024,0.00423200014564726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,16384,0.022054221895005967
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,3584,3584,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,2560,0.03538044293721517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,65536,0.014851555228233337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,65536,16384,0.16639555825127497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,768,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,128,0.014835556348164877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,5120,1536,0.00518844442235099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,3072,0.015397333436542086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,7168,0.0314720008108351
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,6144,0.0122097780307134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,2560,0.023267555567953322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,6144,0.00962933318482505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,8192,0.0262773334980011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,10240,0.016774222254753113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3072,8192,0.025751999682850305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,256,0.01646222174167633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,4096,0.014811555544535318
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,10240,0.03370755579736497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,64,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,3072,768,0.01921066641807556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,10240,512,0.006265777680608962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,6144,0.023948444260491267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,3072,0.06051911248101128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,8192,0.016306666864289176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,8192,0.028408000866572063
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,8192,0.020613332589467365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,256,0.0053119998839166425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2560,256,0.00554666668176651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,3584,0.027688889039887324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,4096,0.007246221933099959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,768,0.014456000592973499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,5120,0.05021244287490845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,4096,0.007568000091446771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,65536,0.08297777838177152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,16384,256,0.004875555634498596
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,2560,0.012697777814335294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,4096,6144,0.04147911071777344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,10240,0.025416889124446448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,2560,0.007368000017272101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,6144,256,0.005886222339338726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,1024,65536,0.0810133351220025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,7168,2560,0.011730666789743634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,1536,0.008972444468074376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,768,0.0038408889538711975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,5120,0.009662222531106737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,512,0.014135110709402295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,6144,0.008645333349704742
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,2048,0.006781333436568578
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,5120,0.12664533985985651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,3584,0.014648889501889547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,64,0.00701155596309238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,256,0.0052355556852287715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,512,0.0041804446114434134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,5120,0.005567111074924469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,5120,0.049321777290768094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,128,0.0058808889653947614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,256,0.0035502223504914176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,256,0.012848888834317526
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,8192,0.024733333124054804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,12288,0.050352000527911715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,7168,0.014684443672498068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,2048,0.008068444828192392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,128,2048,0.006221333311663733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,3072,0.008461332983440822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,512,0.006986666884687211
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2048,1024,0.005922666854328579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,10240,3072,0.01311644415060679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,2560,0.006324444380071428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,128,0.0028453332682450614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,4096,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2560,512,0.003842666745185852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,6144,3584,0.008336000144481659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,10240,0.23648977279663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,8192,3072,0.02306933369901445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,7168,0.01978755493958791
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,8192,5120,0.014142221874660917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,16384,0.008828444613350762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3072,768,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,5120,3584,0.020797333783573575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,65536,0.35673954751756454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,64,0.0034799998005231223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,256,0.006602666858169768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,6144,0.12008799446953668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,1536,0.0064382221963670515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,2048,0.006595555692911148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,8192,0.02258844508065118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,10240,0.013960000541475085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,6144,0.01718844473361969
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,7168,0.022876444790098403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,128,6144,0.0180711117055681
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,8192,0.01313511116637124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,65536,0.15091021855672201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,4096,2560,0.006973333656787872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,65536,0.47575998306274414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,64,0.002961777771512667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,12288,3584,0.015053333507643806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,65536,0.06245155466927422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,10240,0.04894311229387919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,2560,10240,0.0273973329199685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,3072,0.018904000520706177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,12288,5120,0.03148622314135233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,6144,0.019314666589101154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,256,8192,0.030936890178256567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,256,0.0034719999465677473
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,16384,768,0.02695111102528042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,16384,0.025095999240875244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,32,0.008309333688682979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,12288,2048,0.02775466607676612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,65536,0.11177244451310898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,10240,0.09235466851128472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,3072,0.01146755533085929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,6144,2560,0.02013511127895779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,32,0.002773333340883255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1536,6144,0.008990222381220924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,8192,0.11015199952655369
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,768,7168,0.02546399997340308
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,2048,4096,0.007660444411966536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,256,0.020308444897333782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,512,0.011389333340856763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,16384,0.03150399857097202
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,256,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,5120,0.023159111539522808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,512,0.022944000032213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,5120,16384,0.05885066588719686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,65536,10240,0.10828977823257446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,768,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,16384,0.2166373332341512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,7168,0.011683555940786997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,3584,10240,0.026464000344276428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,6144,0.018576888574494254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,1536,0.007639110916190677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,5120,7168,0.012092444631788464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1024,1536,0.008993777963850234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,8192,0.012375999655988483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,32,0.0028311111446883944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,16384,10240,0.03331466515858968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,5120,0.00851022203763326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,16384,0.0281333327293396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,3072,0.07372088564766778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,12288,0.025243555506070454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,10240,0.01535377734237247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,65536,7168,0.1527715524037679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,512,0.0032071111102898917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,7168,0.03478577733039856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,1024,0.029253333806991577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,65536,0.14122133784823948
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,6144,0.020390222469965618
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,1024,0.0620524419678582
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,10240,0.029140445921156142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,128,0.007696888513035244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,64,0.0032355555643637977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,3072,0.006911110960774952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,4096,0.09471822447246975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,256,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,2048,0.005555555638339784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,768,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,32,128,0.0028213332924577924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,4096,5120,0.14552444881863064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,65536,0.07469599776797824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,64,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,10240,0.04288444585270352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,3584,0.005892444401979446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,12288,128,0.015452444553375244
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,3072,0.006511111226346757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,2560,0.00793333351612091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,5120,768,0.02350933353106181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,5120,768,0.017118222183651395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,3584,0.0069919998447100324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,12288,0.050276445017920605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,768,0.016173332929611206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,256,0.005600888695981767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,512,768,0.003920000046491623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3072,1536,0.0069422221018208405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,768,0.005629333357016246
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,128,2560,0.00998222248421775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,8192,2560,0.012469333079126147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,256,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,3584,0.04365155431959364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,256,0.003228444399105178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,64,1024,0.003912888881233004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,1024,0.015903110305468243
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1024,5120,0.008751110898123847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,7168,0.018194667167133756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,2560,0.006949333681000604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,12288,0.017100445098347135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,256,0.005229333208666907
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,768,0.050829334391487964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,10240,0.01182044463025199
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,3584,0.013291555146376291
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,8192,768,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,8192,0.125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,768,0.01754133403301239
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,768,0.005559111220969095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,64,0.005535111245181825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,7168,0.019864888654814828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,7168,3584,0.060106668207380504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,7168,8192,0.018010665973027546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,10240,0.02233155568440755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,768,0.007262222468852997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,256,1024,0.004260444392760594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,32,10240,0.027394667267799377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,64,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,2048,65536,0.10379111104541355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,5120,0.014538667268223233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,65536,1536,0.3457413249545627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,64,0.006231110956933763
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,7168,2560,0.024880889389250014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,4096,0.008361777497662438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,512,8192,0.0107022225856781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,4096,0.007954667011896769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,10240,64,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,7168,10240,0.028140445550282795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,512,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,7168,0.051890667941835195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,16384,0.05412266651789347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,4096,5120,0.023436443673239812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,256,0.007646222081449296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,512,0.03605066736539205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,256,0.013071999781661563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,16384,0.020232889387342665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,8192,0.5740417904324001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,16384,0.020250666472646926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,10240,768,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,10240,0.014092443717850579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,65536,0.06716889142990112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,256,0.00328177772462368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,256,0.007692444655630324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,3584,0.06633422109815809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,7168,10240,0.03125866585307651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,2560,0.006541333264774746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,768,0.005550222264395819
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,768,512,0.003590222034189436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,16384,0.0390897790590922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,4096,0.037061333656311035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,1024,0.017633777525689866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,32,1536,0.010372444159454769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,1536,0.013173333472675748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,5120,0.013061333033773633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,7168,0.019896888070636325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2048,768,0.005568888866239124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,768,0.006647999915811751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,65536,2048,0.20690488815307617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,6144,0.03333510955174764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,2560,0.009711999860074785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,2560,0.04335200124316745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,8192,10240,0.018813333577579923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,2048,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,2048,65536,0.145142224099901
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,3584,0.03167111012670729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,2048,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,128,8192,0.008364444805516137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,1536,0.00700800038046307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,8192,0.013712888790501488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,256,0.03595822387271457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,3584,32,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3072,64,0.019948444432682462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,16384,10240,0.0566142201423645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,10240,8192,0.023062222533755835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,4096,0.014892444014549255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,512,0.007484444313579136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,1536,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,3072,32,0.0069013333155049225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,32,0.005580444302823808
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,1024,0.01612355477280087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,2048,0.007730666961934831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,3584,0.008301333420806462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,12288,5120,0.019371555911170114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,4096,0.22387377421061197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,6144,0.012484444512261285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,256,0.0035457776652442086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,3072,0.04934844374656677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,768,0.007277333074145847
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,1024,0.013198222551080914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,2048,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,4096,0.050557331906424634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,8192,16384,0.02621600031852722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2560,12288,0.08237422174877591
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,4096,1536,0.005971555494599872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,2560,5120,0.010101333260536194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,3072,0.009555555880069733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,4096,0.008022222254011366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,16384,0.4845875634087457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,1024,0.007352888584136963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,2560,0.016010666886965435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,2048,16384,0.021369778447681006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,16384,0.016051555673281353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,10240,0.01645066671901279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,256,0.005407111098368962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,12288,1536,0.010377777947319878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,3584,0.040694223509894475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,2048,0.006735999965005451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,4096,0.3119315571255154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,1024,0.0052426668504873914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,6144,0.01627288924323188
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,4096,8192,0.013528889252079858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1024,2048,0.006222222414281633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,6144,4096,0.027111111415757075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,512,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,3072,0.008037333687146505
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,4096,0.03959111041492886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,16384,0.016513778103722464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,3584,0.006931555353932911
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,128,768,0.003620444486538569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,3584,0.007816889219813878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,8192,0.019875556230545044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3584,10240,0.032277332411872015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,3072,0.07499555746714275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,1536,1536,0.006155555446942647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,6144,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,256,256,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,1024,0.00499199993080563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,32,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,12288,12288,0.04674222071965536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,3072,6144,0.012098666694429187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,6144,65536,0.07407022184795804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,10240,0.027913779020309448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,16384,6144,0.024534222152498033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,7168,0.021879111727078755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,2560,0.014479110638300577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,128,65536,0.025360888904995386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,768,0.0062773335311147906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,768,256,0.006259555617968242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,128,0.007229333122571309
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,512,0.01254666679435306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,1536,0.016622222132152982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,4096,8192,0.015233778291278414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,16384,16384,0.04632444514168633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,12288,0.017064000169436138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,32,0.0086986662613021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,256,6144,0.012088889049159156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,1024,768,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,16384,0.016102222932709586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,7168,0.014564444621404013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,4096,6144,0.0120328888297081
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,128,0.015438222222858004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,2048,0.007998222278224098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,4096,0.007697777615653143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,512,0.013088000317414602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,768,0.005583999885453119
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,4096,0.008291555775536431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,768,128,0.0052826665341854095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,768,256,0.005239999956554837
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,10240,512,0.006264888991912206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,2560,0.048624889718161694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,12288,0.02537155482504103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,3072,0.035375111632876925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,8192,0.06862311230765449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,1024,5120,0.018882667024930317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,1536,0.00794311116139094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,1024,0.011349333657158745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1536,10240,0.012030222349696688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,128,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,65536,0.045180443260404796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,256,32,0.006256000035338932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,5120,0.020817778176731534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,3072,0.06341866652170818
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,1024,0.004184000194072723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,512,8192,0.010336000058386061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,2560,0.01148088855875863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,12288,12288,0.05134222242567274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,12288,0.03780444463094076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,512,0.01276622215906779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,32,0.006070222291681502
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,3584,0.09134399890899658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,256,0.003842666745185852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,128,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,512,0.01649777756796943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2560,1536,0.035624000761244036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,6144,0.008640000389681922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,5120,0.01792533364560869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2048,512,0.006562666760550604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,128,1536,0.0075777777367168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,64,3584,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,1536,256,0.0031840000301599503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,8192,0.016883555385801528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,32,64,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,768,0.009378666679064432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,2048,0.006554666492674086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,1536,256,0.0069439998931354946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,512,768,0.015536889433860779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,2560,0.008393777741326226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,5120,3072,0.14824977185991076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,7168,0.028744889630211726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,6144,0.013442666994200813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,3584,0.00832266691658232
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,1024,1536,0.0046897778908411665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,1536,0.018911111685964797
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,6144,0.030704001585642498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,768,1536,0.014105778601434497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,768,0.005611555443869696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,256,0.02027555637889438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,12288,256,0.017292444904645283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,32,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,12288,0.013110222087966071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,128,0.005952888892756567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,3072,0.007517333659860823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,5120,0.022416889667510986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,12288,0.016516443755891588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,2560,0.007592889169851939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,6144,0.01792355544037289
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,2048,0.011252444651391772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,6144,0.009639999932712978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,4096,32,0.0031351111829280853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,8192,0.01069866700304879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1024,768,0.0232968893316057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,7168,0.029362665282355413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,5120,0.008372444245550368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,768,0.007658667034573025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,256,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,768,0.013517333401574029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,7168,0.022328888376553852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,16384,8192,0.07984266678492229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,768,0.0038524443904558816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,4096,0.007288888924651676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,12288,0.041351109743118286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,6144,0.023825777901543513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,3072,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,32,0.0032097777972618737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,256,0.003527111063400904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,1024,0.005883555445406172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,12288,0.018535999788178336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,4096,0.034367998441060386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,2560,0.007607111500369177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,1024,0.00563822231358952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,128,0.00349866681628757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,7168,0.009699555734793345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,16384,0.013146667016877068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,32,0.003134222287270758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,2560,0.02479466630352868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,4096,256,0.014821334017647637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,7168,0.05065955718358358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,512,4096,0.011723555624485016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,5120,0.009313777916961247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,3584,0.0076044441925154785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,128,0.0029200000895394217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,65536,2560,0.027227555712064106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,768,16384,0.021524444222450256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,2560,0.0069644442862934535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,256,12288,0.01074044406414032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,10240,0.019941333267423842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,128,3584,0.014097778333557976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,6144,0.02590311070283254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,2048,0.02297244469324748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,65536,0.6491973135206435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,12288,0.013175110850069257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,2048,0.007017778025733099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,256,7168,0.0069022224181228215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,3584,0.01316177762216992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,128,1024,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,768,128,0.0034977777136696708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,512,65536,0.05593333641688029
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,768,0.003900444342030419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,12288,0.038338667816585965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,16384,0.13456177711486816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,3584,512,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,10240,0.012308444413873883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,1024,1024,0.013751111096805997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,32,0.002534222271707323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,1024,0.0038568890757030914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,128,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,16384,3072,0.02628533376587762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3584,256,0.005277333160241445
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,1024,0.00998133338159985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,512,16384,0.15310045083363852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,1024,0.00488800017370118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,1024,0.017510222064124215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,2048,2048,0.006813333266311222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,65536,1024,0.0318159990840488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,2560,0.018600889378123812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,4096,6144,0.012410666379663678
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,8192,0.011325333681371478
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,3584,0.10051111380259196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,2560,0.011961778004964193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,5120,0.07449066638946533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2048,6144,0.009669333696365356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,10240,65536,0.14115999804602727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,4096,0.026027555267016094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1024,256,0.0032008888406885993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,256,0.016843555702103507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,4096,0.04855644371774462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,1024,0.00628977765639623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,512,0.014428445034556918
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,8192,0.021003555920388963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,768,0.01643733349111345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,5120,0.00829777783817715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,8192,7168,0.03542666633923849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,256,0.01938311093383365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,65536,0.0893511109881931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,1024,0.01991911066903008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,5120,0.014397333065668741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3072,128,0.014763555592960782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,5120,65536,0.07419555717044406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,768,0.01443822185198466
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,2560,0.008252444366614023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,256,3072,0.007917332980367873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,16384,0.022856000396940444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,10240,5120,0.018805333309703402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,64,0.0028648889727062653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,65536,0.2805466651916504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3072,5120,0.0103511114915212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,256,0.003160888950030009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,8192,12288,0.038727111286587186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,3584,0.04272977842224968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,2048,0.008662222160233391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,1024,0.022200889057583276
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,5120,1024,0.005944889038801193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,2560,512,0.013775111072593264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,768,3584,0.0712053312195672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,5120,0.1697946654425727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,10240,0.017648888958825003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,10240,0.03780622283617655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,4096,0.03874222106403775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,5120,256,0.015554666519165039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,8192,0.04084533452987671
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,10240,0.023716444770495098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,16384,0.4912231233384874
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,12288,7168,0.022562666071785822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,6144,8192,0.017123555143674214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,16384,0.05697155661053128
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,1536,0.004565333326657613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,10240,0.04530844423505995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2048,12288,0.016140444411171805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,4096,65536,2.0368205176459417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,10240,0.1445795562532213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,128,768,0.017877333694034152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,16384,256,0.01754488878779941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,10240,0.0236488895283805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,768,0.014815110299322339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,65536,0.06227999925613403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,2048,4096,0.01239377756913503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,256,0.015444444285498725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,3072,0.0086986662613021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,10240,4096,0.01724177764521705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2560,7168,0.012488889197508493
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,65536,7168,0.1533271074295044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,16384,6144,0.02408355474472046
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,2560,0.007256888680987888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,128,0.004256000121434529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,16384,0.015506666567590503
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,256,4096,0.015830222103330824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,64,0.0029022221763928733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,3584,0.020495999190542433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3072,16384,0.023011555274327595
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,7168,0.009933333430025313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,7168,0.017846221725145977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,512,2048,0.013801777528391944
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,1536,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2048,2048,0.020911999874644812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3072,128,0.010404444403118556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,4096,0.01499466598033905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,3072,0.007287999822033777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,32,0.005632888939645555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,4096,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,1536,0.026552889082166884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,12288,0.041658666398790144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,256,0.039311998420291476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,768,0.005960000058015187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,32,0.0029004443850782183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,12288,10240,0.1943626668718126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,512,0.007110222346252865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,8192,256,0.007148444652557373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,1536,256,0.005320888840489917
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,4096,0.014127999544143677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,768,3584,0.007305777735180325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,256,10240,0.03689510954750909
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,32,0.0028391112056043413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,256,0.017865777015686035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,2048,3072,0.007241778075695038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,10240,0.048672000567118325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,32,0.005247999810510212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,64,128,0.0069066666894488865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2560,12288,0.016167110866970487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,4096,128,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,5120,0.008683555656009251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,512,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,768,0.027802666028340656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2560,6144,0.0658222238222758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,128,0.004850666556093427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,5120,0.005938666562239329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,512,3584,0.006956444846259222
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,3584,12288,0.01761777698993683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,8192,0.01703288820054796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,3072,0.011133333047231039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,12288,0.027638221780459087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,2048,0.005979555762476391
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,10240,0.05788177914089627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2048,7168,0.07493599918153551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1024,1536,0.0046311111913787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,8192,16384,0.050740443997912936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,16384,7168,0.04454133245680067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,3072,0.014352000421947904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,32,5120,0.007291555404663086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,10240,4096,0.07998310857348971
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,7168,64,0.002838222309947014
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,128,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,512,12288,0.04532533221774631
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,7168,0.01626755628320906
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,512,768,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,12288,3072,0.015255111787054272
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2048,2560,0.017872888180944655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,3072,0.020482665962643094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,8192,0.01275022245115704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,4096,10240,0.014763555592960782
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,7168,0.014312000738249885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,3072,4096,0.016181333197487723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1536,16384,0.0303111109468672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,512,0.012446222205956778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,2560,0.05641688903172811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,7168,64,0.003522666792074839
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,5120,8192,0.07544710901048449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,128,2048,0.008681777450773451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,12288,0.01925066610177358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,3072,0.0879475540584988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,12288,7168,0.03596977723969354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,3072,1536,0.014775110615624322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,7168,1024,0.01753866672515869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,128,0.003152888889114062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,768,256,0.005972444597217772
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,768,256,0.0058853332367208265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,256,768,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,2560,5120,0.008919999831252629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,5120,0.015840889679061044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,4096,5120,0.01146755533085929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,1024,0.04755289024776883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,128,1024,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,5120,0.0415653321478102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,10240,32,0.0069102222720781965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,65536,0.05848889218436348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,128,0.014396443963050842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,10240,0.02090399960676829
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,7168,12288,0.03944355580541823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,7168,0.012368000215954252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1536,1536,0.006577777779764599
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,1536,0.00905955582857132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,128,0.014827556080288358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,64,0.0035546666218174826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,5120,6144,0.013094222380055321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,7168,0.015398222539159985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,128,10240,0.012005333271291522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,768,768,0.006246222390068903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2048,64,0.009402666654851701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2048,65536,0.17385866906907824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,4096,0.008285333712895712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,16384,5120,0.0177875558535258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,6144,65536,0.07621422078874376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,768,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,8192,0.8428879843817817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,10240,0.043014221721225314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1024,4096,0.011069333387745751
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,1024,0.02645333276854621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,16384,0.0368755559126536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,4096,0.007659555309348636
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,5120,0.01716444392999013
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,6144,2560,0.01850488947497474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,256,0.003085333233078321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3072,512,0.01585244470172458
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,6144,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,512,0.012817777693271637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,10240,0.33518844180636936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,7168,3072,0.010429333481523724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,3584,0.07842666572994657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,5120,0.017840888765123155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,16384,128,0.0069715554515520734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,5120,0.022056000100241765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,1536,6144,0.019017777509159513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,256,0.005896889087226655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,5120,0.009013333254390294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,32,0.007095110913117726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,64,0.0028115556471877625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,512,0.005548444473081165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,16384,0.01532799998919169
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,3584,0.06050133042865329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,12288,0.014071111877759298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,16384,4096,0.023778667052586872
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,1536,0.026272889640596177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,3584,128,0.0148062225845125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,7168,0.0138568886452251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,4096,6144,0.018665777312384713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2560,128,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,1536,0.006585777633719974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,12288,0.07374133004082574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,5120,6144,0.016586666305859882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,128,32,0.0028364445186323593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,1024,0.006696888970004187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,10240,0.015857777661747403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,16384,3072,0.013603555659453074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,128,2048,0.014826666977670459
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1536,3584,0.032636443773905434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,512,12288,0.0243520008193122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,3072,0.03248000144958496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,12288,0.013770666387346057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3584,16384,0.023538667294714186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,8192,3072,0.027215111586782668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,65536,4096,0.914430194430881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,768,0.015663110547595553
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,256,0.003591999825504091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1024,512,0.011408000356621213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,7168,0.023195554812749226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,128,0.003236444460021125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,1536,0.017887999614079792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,8192,0.010443555812040964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,3072,0.007155555817815993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,768,2048,0.00527022240890397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,1536,0.008430221842394935
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,2048,0.006313777632183499
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,3072,0.03552622265285916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,32,6144,0.007770666645632849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,7168,256,0.004271111140648524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,65536,0.0668462249967787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,2560,0.07886399825414021
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,10240,0.04288088944223192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,6144,0.009706666900051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,7168,1536,0.017254221770498488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,12288,0.015555555621782938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,2048,0.014762666490342883
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,128,12288,0.022290666898091633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,768,0.006072000082996156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,2560,0.024139554964171514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,128,1536,0.0343795551194085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,2560,0.05534222059779697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,8192,0.039098666773902044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,7168,0.011402666568756104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,65536,0.049154665735032826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,128,1024,0.00903733323017756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,6144,0.009016888837019602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,2560,3072,0.020649777518378366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,2560,0.20095111264122856
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,10240,512,0.007686222592989604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,64,3584,0.01068800025516086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,256,3072,0.014522666732470194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,10240,0.052172445588641696
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2048,6144,0.01036177741156684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,5120,256,0.0059075554211934405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,32,0.0038533334930737815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,2048,0.06211644411087036
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,2560,0.01017066670788659
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1536,128,0.006625777731339137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,3584,2560,0.008787555827034844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,3072,0.007400000260935889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1024,5120,0.02816177739037408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1024,2560,0.0068488890926043195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,3584,0.007701333198282454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,8192,16384,0.026330666409598455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,512,0.0335804455810123
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,3584,0.007638221813572778
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,3072,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2560,5120,0.010037333601050908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,1024,0.013458666702111563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,768,7168,0.01032266683048672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,16384,0.02202400068442027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,64,4096,0.007660444411966536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3584,512,0.003928889003064898
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,8192,0.01682222220632765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,512,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,1024,5120,0.008604444563388824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,1536,0.00869066682126787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,768,8192,0.021003555920388963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,1536,0.006011555592219035
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,4096,0.009291555318567488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,10240,7168,0.019996444384257
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,2048,0.016512889001104567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,256,0.005648889061477449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,6144,0.01796444422668881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,10240,3584,0.030606223477257624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,5120,0.07474044296476576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,16384,0.02206666602028741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,2048,0.023005333211686876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,256,2560,0.014800000521871777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3584,65536,0.06676710976494683
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,4096,0.0394720004664527
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,5120,10240,0.026923555466863845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,5120,256,0.005951111101441913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,768,0.003567111161020067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,3584,0.0710808899667528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,16384,0.040531555811564125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,65536,0.11427288585238987
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,7168,32,0.0035866668654812705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,65536,0.06652888986799452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2048,12288,0.013393777940008374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,12288,1.1611519919501412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,5120,0.01758666667673323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,10240,0.012349333200189801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,2048,0.005314666777849197
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1536,7168,0.013550221920013428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,5120,5120,0.014408888088332282
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,3072,0.010967111421955956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2048,2560,0.006156444549560547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,256,0.005807999935415056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,256,0.007079111205206976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,4096,6144,0.01700888905260298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,10240,0.03183111217286851
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,3584,0.00794044468137953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,12288,0.09362577729754978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,16384,0.01574577722284529
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,12288,0.020600888464185927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,2560,256,0.005596444424655702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,3072,16384,0.022224000758594934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,8192,2048,0.07806489202711317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,32,3072,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,7168,0.14712977409362793
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,2560,0.015643555257055495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,1536,0.006782222125265334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,1536,0.0240639994541804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,6144,0.023767999476856653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,10240,0.018040888839297824
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,12288,0.1308177842034234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,512,128,0.005280888742870755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,8192,0.0399217771159278
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,65536,0.08309600088331434
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,4096,0.012656889028019376
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,768,65536,0.25029510921902126
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,1024,6144,0.02071466710832384
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,12288,0.023415111833148535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,5120,0.028604444530275133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,12288,2560,0.04101955559518602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,5120,3072,0.03982222080230713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,5120,0.009848000274764167
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,2048,0.006637333167923822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,3584,0.007904888855086433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,512,12288,0.02369155486424764
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,256,0.015062222878138224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,4096,16384,0.04313155677583483
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,128,32,0.0069644442862934535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,128,256,0.0052897776994440295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,4096,0.008350222474998897
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,1536,5120,0.008375110725561777
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,512,0.017234666479958426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,128,64,0.0028693332440323303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,128,256,0.00630666646692488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,128,0.01568000018596649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,2048,0.008099555969238281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,10240,512,0.018245332770877414
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,3072,0.007674666742483775
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,8192,0.021732444564501446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,65536,0.14690221680535212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,4096,12288,0.04923199945025974
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,768,6144,0.018239110708236694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,1024,2560,0.014820444915029736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,7168,256,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,16384,0.0064533332155810455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,768,65536,0.0280693338976966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,1536,0.006360888895061281
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1536,2048,0.006711111300521427
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,6144,128,0.010228444304731157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3584,5120,0.020818667279349435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,128,0.015112888481881885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,1536,0.22859022352430555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,7168,0.020396444532606337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,65536,0.06934489144219293
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,5120,12288,0.03228444523281521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,5120,3584,0.008115555677149031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,256,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,5120,0.007956444389290279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,1024,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,16384,0.012280888855457306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,8192,0.013763555222087435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,1024,0.005954666684071223
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,8192,1536,0.019239111079110038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,32,2560,0.007296889192528195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,1024,0.0038613333470291565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,768,768,0.0038844446341196695
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,2048,1536,0.0069306666652361555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2560,1024,0.011200889117187925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,6144,64,0.011760888828171624
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,65536,512,0.08391111426883274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,8192,0.02621511121590932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2560,32,0.003559999995761447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,12288,0.0626737806532118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,5120,0.009110222260157267
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,10240,0.14760977692074245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,1024,0.011552888486120435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,65536,1024,0.018609777092933655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,1536,0.006308444258239534
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,256,0.010495111346244812
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,7168,0.14520977603064641
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,16384,16384,0.048391110367245145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1024,1536,0.014455111490355598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,512,0.005543111099137201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,8192,2560,0.024901333782407973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,12288,0.030714667505688135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,8192,64,0.025247111916542053
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,8192,0.011360000405046674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,1024,128,0.0034942221310403612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,2560,0.02405422263675266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,1536,0.007847111258241866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,512,0.02049066623051961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1536,2048,0.007037333316273159
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3584,128,0.006052444378534953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,768,10240,0.009752000371615091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,2048,7168,0.010443555812040964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,65536,256,0.022893332772784766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,256,128,0.005657777604129579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,2560,0.014304000470373364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,32,0.005994666781690385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,16384,0.05031733380423652
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3072,768,0.006907555378145642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,4096,0.012224888636006249
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,7168,0.01222222215599484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,2560,128,0.005432888865470886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,8192,0.031006223625606958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,3072,7168,0.013271999855836233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,7168,10240,0.017669333351982962
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,128,1024,0.005955555372767978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,10240,1536,0.010019555687904358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,12288,0.013906665974193148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,32,512,0.00318755561278926
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,3584,0.011044444309340583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,6144,2048,0.03903111153178745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,3072,0.007302222152551015
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,3584,0.008984888593355814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,256,0.016521778371598985
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,64,0.01072799993885888
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,16384,3072,0.2288444572024875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,4096,6144,0.01238844460911221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,4096,0.01705333259370592
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,6144,3072,0.1781733300950792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,1536,0.005748444547255834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,1024,0.006696888970004187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,8192,0.025666667355431452
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,32,0.006321777900060018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,2048,0.01070577816830741
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,768,0.0074106670088238185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,12288,0.48123375574747723
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,65536,0.25198043717278373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,2048,0.09770933124754165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,1024,128,0.005207111024194294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,3072,512,0.0036391110883818734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,7168,16384,0.18442132737901476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,3072,0.013717333475748697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,12288,0.02552622225549486
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,12288,1024,0.012472000386979846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,7168,3072,0.010772444307804108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,1536,7168,0.026172444224357605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,2048,0.011398221883508893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,10240,12288,0.044334222873051964
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,2560,1536,0.014460444450378418
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,2048,0.0063164445261160536
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,16384,8192,0.818329758114285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,10240,0.03394666645261977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,4096,128,0.022008889251285132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,10240,16384,0.0406675570540958
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2048,512,0.028934223784340754
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,256,8192,0.010677333507272931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,16384,0.023722666833135817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,768,3584,0.01202400028705597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3584,4096,0.008627555436558193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1536,16384,0.019540444016456604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,256,512,0.01958666741847992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,8192,0.092448890209198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,512,3072,0.007080000307824876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,5120,16384,0.04297955499755012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1536,3584,0.015993777248594496
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,512,0.005647999958859549
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,4096,10240,0.1736417743894789
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,3072,0.017265778448846605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,6144,0.021756443712446425
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,3584,0.09937066502041286
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,2560,2048,0.01959288948112064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,32,2560,0.009710222482681274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,12288,5120,0.016303999556435477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,12288,6144,0.028965334097544353
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,65536,768,0.23963822258843315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,4096,768,0.02553155521551768
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,512,0.011735111474990845
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,8192,2048,0.007280000381999546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,32,0.00388711111413108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,12288,256,0.006322666588756774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,6144,128,0.008022222254011366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,2560,3072,0.007299555672539606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,64,64,0.002752888947725296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,6144,0.028424888849258423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,2048,0.006220444622966979
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,5120,4096,0.02204533252451155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,768,0.004593777573770947
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,3072,16384,0.09067200289832221
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,2048,0.006276444428496891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,256,768,0.006695999867386288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,3072,0.02732177740997738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,12288,0.042360888587103955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,256,0.03221688999070062
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,64,0.006673777682913675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,8192,10240,0.028237332900365193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,2560,0.05221422182189094
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,5120,0.011328889264000786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,32,0.0058355554938316345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3072,3072,0.018026666508780587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,65536,128,0.020190222395790946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,512,0.0069511110583941145
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,12288,0.08686400122112697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,4096,2048,0.0169493340783649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,3072,4096,0.007725333174069722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,512,1536,0.014111111561457316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,5120,0.008640000389681922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,12288,10240,0.02660177813635932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,16384,1536,0.03020355436537001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,3584,65536,0.13577511575486925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,512,0.003922666526503033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,12288,768,0.0070017774899800616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,1024,0.08154666423797607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,16384,0.04518577787611219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,2048,0.008947555389669206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,1536,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,8192,0.025596444805463154
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,8192,0.01311377767059538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,3584,0.04271466533342997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,3072,0.014560000763999091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3584,768,0.005577777822812398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,65536,2048,0.22444356812371147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3584,32,0.003549333247873518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,1024,0.00703022215101454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,4096,6144,0.023042667243215773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,256,1536,0.004568888909286923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,8192,16384,0.028299556838141546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,2048,0.04631466666857401
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2560,1024,0.004550222307443619
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,6144,0.03745511174201965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,3072,0.008637333081828224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,7168,6144,0.014735110931926303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,10240,6144,0.03535022338231405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,8192,4096,0.010758221977286868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,4096,128,0.002879111096262932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,1536,0.006567111031876669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2048,65536,0.06071022484037611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,7168,256,0.017196445001496207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,2048,0.008791999684439765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,3584,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,3584,0.017456000049908955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,512,8192,0.013229333692126803
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,32,1536,0.004872888740566042
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,2048,0.028725332683987085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,16384,512,0.0069120000633928515
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,8192,1024,0.009100444614887238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,8192,0.0905111100938585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,768,7168,0.06707377566231622
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,128,2048,0.010289777484205034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,2048,0.006634666687912411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,5120,512,0.012784000072214337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,5120,7168,0.01348622226052814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,7168,32,0.0032311110860771606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,10240,512,0.009712888962692684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,8192,0.0266942216290368
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,3072,0.007374222079912822
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,8192,0.018262222409248352
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,3072,0.014136888914638095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,6144,0.11728355619642471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,768,0.016159110599093966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,7168,6144,0.014759999182489185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3072,1024,0.006801777829726537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,128,0.005707555760939916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,128,0.005315555466545953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,7168,128,0.020273778173658583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,3072,65536,0.09528888596428765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,6144,5120,0.015602666470739575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,8192,0.026522666215896606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,512,0.003588444242874781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1536,3584,0.08407288789749146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,10240,0.013914666242069669
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,512,0.003912000192536248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,3584,0.007019555403126611
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,256,1024,0.003927111211750242
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,2048,8192,0.02641600039270189
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,10240,0.016311999824311998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,6144,0.037596444288889565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,128,0.013417777915795645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,12288,0.01036000003417333
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,64,64,0.0028320000403457214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,4096,0.019274666905403137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,4096,0.00794666674402025
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,16384,2560,0.013139555851618448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,2048,0.006284444282452266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,65536,0.07512266768349542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,6144,6144,0.01825777855184343
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3072,1536,0.015574221809705099
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,32,2048,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,4096,5120,0.018368000785509746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,6144,512,0.005680000202523337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,16384,1536,0.013250666360060373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,7168,768,0.061273780133989125
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,128,0.0032346666687064697
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,256,0.0035573335157500375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,6144,0.013329777452680798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,65536,0.1987182166841295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,3584,2048,0.015581332974963717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,6144,3584,0.01036622209681405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,8192,0.010448888772063784
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,512,0.007977777885066139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,16384,4096,0.01904000010755327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,7168,0.010139555566840703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,1536,0.00598933340774642
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,256,16384,0.009373333719041612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,5120,0.009301332963837517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,768,128,0.005865777946180767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,4096,64,0.003265777809752358
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,4096,0.018591110905011494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,256,0.014882667197121514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,1024,3584,0.016318221886952717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,256,0.006002666635645761
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,8192,0.10394489102893406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3584,3584,0.011454222102959951
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3072,6144,0.17345421844058564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,128,768,0.006025777922736273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,2560,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,1536,0.0063475556671619415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3072,5120,0.008286221987671321
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,512,7168,0.009400889277458191
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,4096,0.007635555333561367
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,7168,512,0.00565155554148886
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,256,512,0.005299555758635203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,4096,0.007654222349325816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,7168,6144,0.03802311089303758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,768,0.00563733321097162
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,8192,0.13732533984714085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,1536,0.008382221890820397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1024,8192,0.010038221875826517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,2560,0.009893333746327294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,3584,0.007975111405054728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,3584,128,0.015088000231319003
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,64,0.005535111245181825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,64,16384,0.015327110886573792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,32,0.012488000094890594
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,4096,0.019753777318530612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1024,10240,0.034318221939934626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,12288,8192,0.03756177756521437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,2560,32,0.00832444429397583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3072,8192,0.013796444568369122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,6144,8192,0.02125777800877889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,256,0.0031804444475306403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,65536,128,0.025235555238193933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,2048,0.006589333216349284
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,512,64,0.007034666836261749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,256,0.009455111291673448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,32,64,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,3584,3584,0.008021333151393467
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,256,4096,0.007763555480374231
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,1536,1024,0.02086488902568817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,2048,0.0831306642956204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,10240,16384,0.05924533473120796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,7168,0.034396443102094866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,10240,0.07795555724038018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,6144,2560,0.011050666371981302
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,16384,0.02816177739037408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3072,16384,0.020624000165197585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,32,2048,0.015767999821239047
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,5120,0.008986666798591614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,16384,0.034550223085615374
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,4096,0.016221332881185744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,128,0.005960888746711943
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3072,16384,0.03277333246337043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,4096,0.016177778442700703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,768,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3584,3072,0.008017777568764156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,768,6144,0.008995555341243744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,1536,0.012867555850081973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,1024,0.028173334068722192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,5120,0.009752888646390703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,256,128,0.006657777975002925
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,512,0.05772799915737576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,8192,5120,0.032060444355010986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,4096,3072,0.010430222584141625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,4096,1024,0.01605688863330417
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2048,3584,0.023774221539497375
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,7168,0.01723288827472263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,256,768,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,256,7168,0.016343999240133498
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,256,0.0052844443255000645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,256,1024,0.01586577792962392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,16384,10240,0.036510222487979464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,10240,3072,0.02640177806218465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,7168,0.018545778261290658
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,4096,4096,0.010732444624106089
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,5120,0.008407110969225565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,65536,0.06268977456622653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,128,512,0.005372444374693765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,16384,0.04276000128851997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,128,512,0.005561777700980504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,65536,64,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1024,5120,0.009026666482289633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,4096,4096,0.020184000333150227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,10240,5120,0.47289689381917316
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2048,3584,0.012030222349696688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,10240,0.07023555702633329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,3584,1536,0.015587555037604438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,32,0.0025946667624844448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,2048,0.04666399955749512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,3584,65536,2.5080479515923395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,32,0.03464355402522617
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,6144,3584,0.012471111284361945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,6144,1536,0.018979556030697294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,16384,8192,0.03710133168432448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,3584,0.00757244477669398
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,5120,16384,0.03669866588380601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,6144,0.028385778268178303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1024,6144,0.018559111489189994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,512,7168,0.020786666207843356
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,1024,0.027422222826215956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,3584,768,0.018906666172875296
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,256,64,0.005668444352017508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,2560,0.0069679998689227635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,2048,0.005381333331267039
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,10240,768,0.008768888811270395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,7168,0.015074667003419666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,5120,10240,0.0307626658015781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,2048,0.019104889697498746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,12288,512,0.01927911076280806
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,12288,1.5026239818996852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,512,0.005593777944644292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,1024,16384,0.015817777978049386
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,16384,768,0.01833422150876787
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,12288,2560,0.01905333333545261
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,1024,0.02574399941497379
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,3584,0.017824000782436795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,5120,8192,0.019872888922691345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,32,0.008737777670224508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,6144,0.09671911266114976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,768,0.024777778320842322
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,10240,128,0.0063324446479479475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,256,2560,0.006744888921578725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,256,256,0.003232888877391815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,3584,10240,0.01622488929165734
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,2560,0.008311111066076491
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,3072,2048,0.006966222491529252
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,2048,0.0069013333155049225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,8192,0.013591999808947245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,10240,0.012108444339699216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,2048,0.013988445202509562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,16384,2560,0.012652444342772165
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,8192,16384,0.05774933099746704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,16384,6144,0.023200000325838726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,4096,128,0.005929777605666055
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,10240,256,0.04310489032003614
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,128,1536,0.004920888692140579
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,128,128,0.010991111397743225
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,4096,7168,0.012822222378518848
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1536,128,0.005218666460778978
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,1536,0.007265778051482306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,1536,0.00423911131090588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,1536,256,0.005910222315125995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,1024,0.006015111174848344
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,1024,0.005601777798599667
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3584,1536,0.006271111054552927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,5120,0.0387733346886105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,256,0.006320000108745363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,1536,0.011623111036088733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,256,8192,0.029747555653254192
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,1536,0.00738933351304796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,64,2048,0.006810666786299811
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,256,0.015226667126019796
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,4096,512,0.006298666612969504
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,2048,0.006256888724035687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,65536,16384,0.6578293376498753
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,6144,0.011393778026103973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,512,3584,0.01622577839427524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,1536,2560,0.007661333514584436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,256,0.006639110959238476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,65536,8192,0.2559013366699219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,65536,3.7450425889756946
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,10240,0.052526222334967725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,128,0.006236444330877728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,128,256,0.024329778220918443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,768,10240,0.036072889963785805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3072,16384,0.020625778370433383
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,1536,512,0.014547554983033074
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,8192,6144,0.02067733307679494
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,2560,768,0.01910666624704997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,768,0.011060444845093621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,8192,0.3047288788689507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,256,3584,0.014521777629852295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,4096,5120,0.011060444845093621
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,8192,0.006698666761318843
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,4096,10240,0.029008888536029395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,3584,0.0074480002125104266
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,65536,768,0.016012445092201233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,512,0.011738667057620155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,256,16384,0.03762311074468825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,7168,128,0.005640888793600931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,8192,0.020928889513015747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,65536,5120,1.0994097391764324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,4096,2560,0.006940444310506185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,256,7168,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,65536,5120,0.05581866370307075
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,2560,0.006603555546866522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,128,0.024558222956127588
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,4096,16384,0.24460355440775552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,2048,64,0.0029031110720502008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,1024,0.027053332991070215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,1536,0.02751200066672431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,5120,0.008716444174448649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,1536,0.006280000011126201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,12288,0.02810488806830512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,10240,3072,0.013950222068362765
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,4096,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,512,768,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,256,768,0.01593955523437924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,256,0.003588444242874781
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,768,1024,0.01350844403107961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,512,0.0053395554423332214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,65536,0.05992799997329712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,512,7168,0.007977777885066139
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3584,10240,0.016003555721706815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,1024,2048,0.008781332936551835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,128,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2048,128,0.01349955548842748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,6144,8192,0.02382844520939721
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,256,0.018036444981892902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,768,0.007358222372002072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,768,6144,0.009730666875839233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,128,1024,0.004033777862787247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,512,8192,0.03056000007523431
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,5120,0.0069875551594628235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,65536,10240,0.18376978238423666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,5120,0.018954666124449838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,3072,1536,0.019353777170181274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,10240,7168,0.062479999330308705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,6144,0.008658666577604082
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,4096,8192,0.01389422184891171
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,768,0.00573955559068256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,5120,32,0.0035742223262786865
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,64,10240,0.0322693321439955
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,64,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,6144,6144,0.23959199587504068
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,5120,8192,0.031742221779293485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2048,64,0.005585777676767773
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,768,3584,0.007698666718271043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,1024,768,0.03188177943229675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,3584,768,0.017628444565667045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,10240,0.0188417782386144
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,12288,0.00692622239391009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,16384,10240,0.24003556039598253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,512,7168,0.010055999788973067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,512,65536,0.09919110933939616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,65536,256,0.05998488929536608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,768,3584,0.015190222197108798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,7168,768,0.005974222388532426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,768,0.02205955485502879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,768,8192,0.07111555337905884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,768,8192,0.010370666782061258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,128,0.006613333192136552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,5120,0.06578310992982653
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,768,0.007332444190979004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,5120,12288,0.026405332816971674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,512,8192,0.01071111112833023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,8192,2048,0.009018667042255402
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,2048,0.02068355513943566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,8192,128,0.014837332897716098
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,128,10240,0.012514666550689273
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2048,5120,0.017645332548353408
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,2560,256,0.0052888890107472735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,512,0.005419555637571547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,768,512,0.01759377784199185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,65536,1024,0.06195733282301161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2560,5120,0.01161688897344801
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,512,2560,0.014176888598336114
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,768,0.005894222193294101
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,1536,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,1536,0.021464889248212177
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,16384,0.01331822243001726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,1024,2560,0.007311999797821045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,5120,0.008999110923873054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,12288,768,0.007272000114123027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3584,16384,0.03273955649799771
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,512,512,0.003185777821474605
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,16384,768,0.013618666264745923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,7168,256,0.016885333591037326
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,10240,0.05681066380606758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,2048,0.008770667016506195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,768,0.01681155628628201
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,128,0.0028373334142896864
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,3584,0.007650666766696506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,32,256,0.0052764444715446895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,5120,0.025399999486075506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,256,8192,0.022008889251285132
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,2560,256,0.006897777732875612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,512,10240,0.012515555653307172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,10240,0.7592506408691406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1024,7168,0.010286221901575724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,128,3072,0.014683556225564746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,5120,0.015146666102939181
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,1024,8192,0.010391111175219217
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,768,6144,0.007887110941939885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,4096,0.0773786637518141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,16384,768,0.008055999875068665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,6144,256,0.005461333526505365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,7168,0.010081777969996134
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,6144,0.013824889229403602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,6144,10240,0.0379342238108317
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,512,7168,0.01903377804491255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,16384,0.007721777591440413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2048,32,0.0069351109365622205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,128,12288,0.022405332989162866
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,8192,65536,0.08458489179611206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,1024,7168,0.010006222460005019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,65536,5120,0.043903999858432345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,2560,7168,0.18862399790022108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1024,2048,0.006919999917348226
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,512,16384,0.02017155620786879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,32,0.006705777926577463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,12288,0.013368888861603208
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,128,64,0.003014222201373842
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,65536,0.0843013326327006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,5120,5120,0.023007111416922674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,512,0.0033777778347333274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,6144,512,0.016581333345837064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3584,768,0.006241777704821692
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,10240,0.04424800144301521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,4096,4096,0.011587555209795633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,12288,6144,0.11381067170037164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,512,4096,0.007264888948864407
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,1536,1024,0.00462400002612008
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,128,2560,0.007155555817815993
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,8192,3584,0.06256888972388373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,128,3072,0.013564444250530668
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2560,10240,0.024922667278183833
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,768,65536,0.06240444713168674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,32,0.30224622620476616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,256,0.006018666757477655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,6144,512,0.003910222401221593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,1024,0.012120000190205045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2048,10240,0.020304888486862183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,2048,0.0062977779242727495
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,10240,1536,0.03870133227772183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,3072,12288,0.17545778221554229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,12288,0.02919644448492262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,512,10240,0.0225608895222346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,256,7168,0.006663110935025745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,8192,0.014508444401952954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2560,768,0.00629066675901413
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,10240,16384,0.05154666635725233
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,16384,0.01902577777703603
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1024,2560,0.005590222362014983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,128,0.005227555417352253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,16384,0.02833777666091919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,7168,1024,0.00981866651111179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,12288,256,0.00328533330725299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,12288,0.013077333569526672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,4096,1024,0.006737777756320105
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,128,0.0032213332338465583
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,6144,0.029097778929604426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,1024,768,0.0069546666410234236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,3584,0.010967999696731567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,10240,2560,0.013740444348918067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,65536,512,0.014833778142929077
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,256,65536,0.38280354605780703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,1536,65536,0.6564915445115831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,2560,0.010444444086816577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,256,0.0041831110914548235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,7168,10240,0.03575911124547323
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,7168,16384,0.02464888824356927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,1536,64,0.005560889012283749
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,8192,16384,0.026388444834285315
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,7168,128,0.009648888475365108
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,8192,0.19362221823798287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,128,1536,0.012808889150619507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,2560,0.00999022192425198
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,10240,16384,0.057553774780697294
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,768,0.005934222290913264
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,65536,0.0625857777065701
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3584,10240,0.02739022175470988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,16384,128,0.006648889018429651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,32,2560,0.0058746664888328975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,512,3584,0.007648000286685095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,8192,65536,0.1098524464501275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,768,64,0.00658044425977601
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,4096,128,0.019672888848516677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,64,2560,0.007335999773608313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,512,2560,0.005952000204059813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,2048,65536,0.06374577681223552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,1536,0.013071111506885953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,8192,0.02626400027010176
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,1024,16384,0.03288977675967746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,6144,0.013434666726324292
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,12288,12288,0.06583999925189547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,512,768,0.0038906666967603895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,2560,7168,0.014834667245546976
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,4096,0.022520000735918682
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,32,0.005579555614127054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,2560,1536,0.014838222000333997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,512,16384,0.012873777912722694
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,3584,1536,0.0069573331210348345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,768,0.020781333247820537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,128,2048,0.006678222368160884
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,7168,65536,0.09772177537282307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,2560,0.008332444561852349
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,1024,0.01348977784315745
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,1024,0.027685332629415724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,4096,7168,0.01891377733813392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,12288,8192,0.31910755899217397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,3072,0.029042667812771265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,12288,0.07937422063615587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,3072,0.007647111184067196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,2560,0.0333555539449056
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,512,0.005592000153329637
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,2048,128,0.0031724443866146933
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,3072,7168,0.013589333328935834
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,5120,0.032829334338506065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,12288,4096,0.01538755496342977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,3072,8192,0.014904000692897372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,3584,5120,0.027520888381534155
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,2560,6144,0.00904977818330129
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,12288,1024,0.007387555307812161
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,12288,0.03106488784154256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,65536,0.02637600070900387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,128,0.006917333437336816
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,65536,32,0.005631999837027655
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,16384,0.03435733252101474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,10240,5120,0.020090666082170274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3584,512,0.008377778033415476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,64,0.0032266666077905228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,12288,7168,0.05292000042067634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,10240,2048,0.009088888764381409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,10240,64,0.01755288905567593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,16384,6144,0.0399253335263994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,8192,6144,0.01870755520131853
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,4096,6144,0.02254399988386366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,1536,2560,0.01811733345190684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,12288,256,0.006279110908508301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,2560,4096,0.01757688820362091
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,6144,256,0.005895111295912001
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,12288,12288,0.11853510803646511
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2560,16384,0.022797332869635686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,64,6144,0.006296000132958095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,65536,256,0.012934221989578672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,2048,8192,0.010935111178292168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,8192,2560,0.010055111514197456
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,2048,32,0.0029315555261241067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,3072,16384,0.21484000153011748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,8192,16384,0.05279111199908786
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,10240,0.029743999242782593
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,3584,0.017303110824690927
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,64,1024,0.005991999887757831
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,8192,1024,0.023777777949968975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,256,6144,0.030604445272021826
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,8192,768,0.005202666752868229
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,65536,1024,0.04455022348297966
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,10240,1024,0.010206222534179688
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,1536,0.006952889263629913
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2560,2048,0.03587822119394938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,768,1536,0.006599111275540457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,6144,4096,0.025377778543366328
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,12288,1024,0.008718222379684448
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,64,12288,0.014103111293580798
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,12288,4096,0.052654223309622876
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2560,256,0.005621333503060871
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,2560,0.021105777886178758
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,512,1024,0.013071111506885953
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,256,6144,0.018530666828155518
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,64,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,2560,1024,0.004729777574539185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,256,32,0.006679111056857639
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,128,3072,0.007280888656775157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,512,0.0032355555643637977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,5120,65536,0.07458400064044528
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,128,256,0.0060159998635451
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,7168,32,0.006251555350091722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,128,0.014874666929244995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,8192,0.014297778407732645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,16384,32,0.0038728887836138406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,2560,768,0.016837333639462788
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,7168,0.023541332946883306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,4096,2048,0.007627555893527136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,65536,0.06072533130645752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,16384,1536,0.02711733347839779
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,5120,0.030712889300452337
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,3072,256,0.005464000006516774
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,4096,1024,0.01614488826857673
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,1024,0.006253333141406377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2560,6144,0.01108977778090371
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,16384,2560,0.016163556112183463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,7168,0.019655999210145738
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,5120,1024,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,1024,10240,0.12484355767567952
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,12288,12288,0.0362026658323076
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,3072,16384,0.021162667208247717
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,512,128,0.0028942223224374983
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,5120,10240,0.015312888556056552
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,2048,32,0.005628444254398346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,768,0.003536888708670934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,3072,10240,0.039826667971081205
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,512,2560,0.005912888795137405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,4096,5120,0.18833866384294298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,6144,5120,0.015022221538755627
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,3584,12288,0.017510222064124215
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,16384,16384,0.04572977622350057
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,256,0.011360888679822287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,2560,1536,0.006274666637182236
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,3072,1024,0.004214222232500712
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1536,512,0.0035662220584021676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,64,128,0.002847111059559716
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,3072,0.020177778270509508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3584,65536,0.10078489118152195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,8192,32,0.0028702221396896574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,6144,6144,0.025518221987618342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,2048,0.006656000183688269
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,10240,2560,0.09719644652472602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,3072,3072,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,2560,0.007278222176763747
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,4096,3072,0.00740799970097012
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,3584,6144,0.009647111097971598
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,16384,10240,0.0558755530251397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,16384,0.028858668274349634
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,2560,65536,0.14489332834879556
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,32,768,0.006063111126422882
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,512,256,0.0032115555885765287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,65536,256,0.008678221868144141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,512,12288,0.013187555803192986
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,16384,6144,0.04134755664401584
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,512,0.012548444171746572
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,5120,0.022639999787012737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,512,1536,0.005008888741334279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,768,0.015309333801269531
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,768,10240,0.011798222031858234
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,8192,6144,0.014004444082578024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,16384,2048,0.015243555108706156
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,8192,10240,0.0337413317627377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,3072,4096,0.12212177117665608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,32,12288,0.006643555644485686
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,3072,0.012983111043771109
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,10240,2048,0.011824888487656912
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,3072,32,0.002854222224818336
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,8192,3584,0.011698666546079846
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,1536,1536,0.015989333391189575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,4096,0.012455110748608908
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,5120,12288,0.036475555764304265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,3072,0.02696000039577484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,128,32,0.005605333381228977
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,12288,6144,0.0754844413863288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,12288,3072,0.01737333337465922
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,32,0.006005333529578314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,5120,3072,0.02012888921631707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,2048,0.014146667387750415
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,16384,16384,1.1935813691880968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,768,512,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,1024,256,0.027935110860400732
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,65536,7168,0.0869999991522895
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1536,256,0.006959999601046245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,65536,1024,0.01827911039193471
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,12288,8192,0.02347911066479153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,3584,64,0.00592533333433999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,16384,65536,0.196051557858785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,128,2048,0.013456000222100152
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,6144,3072,0.03160533308982849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,32,0.0058959999846087555
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1024,8192,0.026504889130592346
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,3584,0.007167999943097432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,8192,256,0.008455110920800103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,512,1024,0.015595555305480957
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,10240,5120,0.02924622098604838
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,12288,2048,0.013354666531085968
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,7168,3584,0.11667911211649577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,1536,65536,0.4698790974087185
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,256,3072,0.01554666625128852
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,10240,1024,0.008839999636014303
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,1024,0.005967111223273807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,12288,1024,0.03498311175240411
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,768,128,0.01068800025516086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,7168,6144,0.014895111322402954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,5120,12288,0.04245511028501722
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,1024,0.016499555773205228
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,5120,0.011126221881972419
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,32,0.005981333139869902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,3072,2048,0.0069724445541699724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,512,3072,0.007306666837798224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,512,0.014882667197121514
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,3072,0.00701155596309238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,1536,4096,0.008627555436558193
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,8192,0.016166221764352586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,2048,0.015174221661355762
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,10240,0.012512000070677863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1536,768,0.01759822169939677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,4096,128,0.003516444315512975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,128,7168,0.006405333264006509
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,2560,8192,0.018779555956522625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,10240,0.013441777891582914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,4096,512,0.0063777777055899305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,7168,0.25981065962049693
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,12288,512,0.006616888774765863
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,8192,0.10681511296166314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,2048,0.027126222848892212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,128,0.00600355573826366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,12288,12288,0.22120266490512422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2560,7168,0.01144444445768992
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,768,4096,0.016186666157510545
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,128,2048,0.013336000343163809
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,10240,512,0.037956443097856306
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,10240,3584,0.03617599937650893
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,256,16384,0.015454222758611044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,256,256,0.0052471111218134565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,2560,0.0064186664919058485
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,32,64,0.0026604444202449587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,65536,1024,0.02863111098607381
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,10240,6144,0.02998577886157566
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,65536,0.008728000024954477
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,8192,1536,0.00792711145348019
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,4096,8192,0.02515022291077508
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,8192,256,0.046557333734300405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,3072,768,0.005587555468082428
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,2560,1536,0.014514666464593677
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,10240,2048,0.008859555754396651
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,64,8192,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,16384,0.2626230981614855
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,8192,0.01961866683430142
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2560,16384,0.11899467309315999
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,65536,1536,0.06700088580449422
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,16384,1024,0.008628444539176093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,2560,12288,0.028445333242416382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,64,0.0029395555870400537
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,768,0.003556444413132138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,2048,1536,0.012072888513406118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,6144,0.009741333623727163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,3584,3584,0.007992888490358988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,512,0.005942222144868638
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,7168,8192,0.03162311183081733
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3584,3584,0.007993777592976889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,7168,0.04284444451332092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,6144,16384,0.11190756162007649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,16384,64,0.0035991109907627106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,768,10240,0.012845333251688214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,12288,2048,0.010503110786279043
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,2048,12288,0.01651022169325087
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,1024,12288,0.012489777472284106
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,10240,3072,0.04413777920934889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,768,7168,0.02330400049686432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,65536,0.13174843788146973
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,512,65536,0.06330666939417522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,65536,0.048136889934539795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,8192,0.01865333318710327
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,12288,2560,0.012151111331250934
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,2048,5120,0.009741333623727163
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,1024,0.00591466658645206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,4096,128,0.014503111441930136
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3072,4096,0.01352266636159685
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,7168,6144,0.011875555747085147
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,3584,2048,0.0069395556218094295
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1536,2560,0.006254222244024277
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,2048,0.0251511103577084
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,32,128,0.0028355556229750314
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,1536,0.02159999973244137
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,3584,7168,0.01945600079165565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,1536,32,0.0028435554769304064
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,64,128,0.006605333338181178
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,65536,5120,0.052321778403388135
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,10240,2048,0.02868000004026625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,10240,0.018559111489189994
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1536,4096,0.021299555897712708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,65536,2048,0.0724782215224372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,64,2048,0.0069226668112807805
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,4096,6144,0.01241422196229299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,8192,12288,0.033731556600994535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,5120,0.02825333343611823
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,32,768,0.005909333212508096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,4096,0.007965332931942409
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,8192,2560,0.08680088652504815
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,1024,2048,0.014426666829321118
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,64,1536,0.004568888909286923
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,32,16384,0.006615110983451207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,1536,12288,0.016640888320075143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,32,5120,0.010471111370457543
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3584,7168,0.013178666432698568
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,10240,0.13011022408803305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,65536,2048,0.02791822287771437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,7168,65536,0.07400977611541748
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,65536,768,0.041265779071384005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,10240,10240,0.0351475543446011
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,8192,16384,0.03899733225504557
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,1024,6144,0.008996444443861643
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,2048,3584,0.016778666112158034
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,16384,4.293889787462023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,1024,2048,0.009389333426952362
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,2048,3072,0.01074044406414032
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3584,3072,0.02606933315594991
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,64,12288,0.011728000309732227
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,5120,1536,0.008919999831252629
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,768,12288,0.01891288823551602
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,12288,7168,0.03597955571280585
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,16384,64,0.003936000168323517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,65536,64,0.014359111587206522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,8192,65536,0.09094133641984727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2048,8192,0.011075555450386472
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,8192,0.014873777826627096
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,768,16384,0.016207999653286405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,4096,32,0.003142222141226133
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,65536,5120,3.1191217634412975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,2048,1024,0.01406222250726488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,65536,128,0.018947554959191214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,12288,2048,0.01422755585776435
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,2048,0.006596444381607904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,8192,3584,0.011600889265537262
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,8192,0.021669333179791767
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,8192,6144,0.03595822387271457
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,80,6144,2560,0.007328000333574083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,12288,7168,0.03683377636803521
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,5120,7168,0.02516888909869724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,1024,2048,0.014160000615649752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,1536,4096,0.008027555214034187
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,6144,16384,0.0384168889787462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,3584,1536,0.006351999938488007
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,256,8192,0.019967110620604623
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,3072,3072,0.007287110719415877
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,256,5120,0.01016977760526869
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,5120,0.010067555639478896
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,4096,0.006686222222116258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,16384,3584,0.038302222887674965
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,16384,4096,0.372691551844279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,256,7168,0.0068986668354935115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,512,4096,0.07229333453708224
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,4096,1024,0.006272888845867581
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,768,0.004187555362780889
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,2048,0.014526221487257214
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,1024,768,0.007612444460391998
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,256,3072,0.007282666862010956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,256,65536,0.018205333087179396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,12288,12288,0.042075554529825844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,10240,8192,0.024720888998773363
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,512,65536,0.05525866813129849
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,2560,5120,0.04298311140802172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,256,3072,0.006953777538405524
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,6144,6144,0.018423111902342904
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,1024,65536,0.060312887032826744
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1024,3072,0.009331555830107795
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,12288,1536,0.07879910866419475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2560,6144,0.011523555550310345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,5120,3584,0.009678222239017487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,5120,1024,0.015538665983412
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,4096,1024,0.015627556376987033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,512,128,0.0032960000551409195
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,7168,0.014281777871979607
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,4096,64,0.0031893334041039148
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,6144,3584,0.01017599966790941
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,128,7168,0.019679110911157396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,64,256,0.0039057777159743835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,1536,0.010466666685210334
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,3584,0.009356444080670675
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,3072,16384,0.035094221433003746
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,256,128,0.01311911145846049
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,256,12288,0.013203555511103736
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,6144,3584,0.008078222473462423
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,12288,7168,0.034952001439200506
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,2560,0.018931556079122756
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,2560,768,0.013771555489963956
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,1536,4096,0.008030222521887885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2,7168,2048,0.006675555474228329
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3072,3584,0.011374221907721626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,128,2048,0.005943111247486538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,10240,3072,0.012881778180599213
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,256,6144,0.009172444542249044
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,2048,10240,0.015660444895426433
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,6144,32,0.00397155558069547
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,65536,5120,1.064361784193251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,256,512,0.01687022215790219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,6144,0.018036444981892902
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1536,10240,0.016947555873129103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,1536,32,0.0028897778441508612
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,7168,16384,0.025895110434956018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,64,0.0029164445069101122
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,3584,768,0.005995555470387141
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,256,0.017264000243610807
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,8192,3584,0.012850667039553324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,12288,0.016912889149453905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,1536,10240,0.01145511120557785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,2048,1024,0.016191111670600045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,768,0.003583111282851961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2048,2560,0.007315555380450354
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,1024,6144,0.035051554441452026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,8192,3584,0.03132088979085287
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,2048,12288,0.01738133364253574
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,16384,5120,0.053123556905322604
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,12288,3072,0.02715377840730879
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,3584,2560,0.04419555597835117
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,4096,2048,0.02405599918630388
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,1536,0.020981333321995203
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,128,4096,0.005581333405441708
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,64,0.028512001037597656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,3072,4096,0.017986666825082567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,12288,65536,0.14823910925123426
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,2048,4096,0.010024000373151567
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1024,16384,1536,0.018726223044925265
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,8192,0.03001333276430766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,1024,2048,0.006623999940024481
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,7168,0.058155556519826256
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,12288,64,0.03967466619279649
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4096,512,10240,0.03384622269206577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,6144,1024,0.016918222109476726
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,768,0.014583999911944071
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,1536,7168,0.010059555371602377
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,16384,1536,0.06080444653828939
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,7168,768,0.004243555582231945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,65536,7168,0.12433866659800212
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,16384,0.017895999881956313
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,6144,0.018601778480741713
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,6144,512,0.007299555672539606
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,2048,2560,0.0074275558193524676
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,3072,256,0.0036764442920684814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,10240,10240,0.041308446062935725
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,5120,5120,0.01556622154182858
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,5120,1024,0.009921777579519484
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,12288,10240,0.06858666737874348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,512,1024,0.013134222063753339
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,65536,128,0.010048000348938836
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,8192,128,0.01864888932969835
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,6144,0.027829332484139338
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,6144,3584,0.1598844395743476
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,5120,256,0.0038613333470291565
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,64,65536,0.010048888623714447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,384,3584,7168,0.018779555956522625
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,1536,3072,0.015488000379668342
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,12288,768,0.006618666566080517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,10240,0.23370488484700522
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,768,1536,0.01699555582470364
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,7168,64,0.0032195556494924757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,1024,2048,0.017287110288937885
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,256,7168,0.009387555221716562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,1024,3584,0.00666222224632899
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,1024,1536,0.004885333279768626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,65536,16384,1.7689972983466253
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,16384,64,0.01751733322938283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,768,3584,0.013069333301650153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,5120,1536,0.1086417767736647
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,64,64,0.00424177779091729
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,64,12288,0.013119999733236102
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,1024,2048,0.02240444388654497
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,7168,512,0.01648355523745219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,5120,7168,0.027479999595218237
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,8192,7168,0.017254221770498488
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,2048,0.006607999818192587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,768,2560,0.007319110963079665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,4096,0.007985778152942657
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,5120,5120,0.016360888878504436
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,1536,1536,0.004957333207130432
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,64,4096,0.00796888851457172
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,10240,3584,0.013577777478430005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,5120,8192,0.05883466535144382
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,3072,16384,0.04863644308514065
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,16384,1024,0.00814933329820633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,512,4096,0.03115288747681512
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,3072,512,0.00327822214199437
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,2048,1024,0.006333333336644703
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,1024,0.006133333262470033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,512,10240,0.032610668076409235
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,2048,7168,0.013720889058378009
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,7168,1536,0.017647110753589206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,512,16384,0.01071466671095954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,6144,2048,0.020831111404630873
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,65536,0.16056711143917507
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,2560,32,0.0028284444577164124
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,3072,2048,0.006584888945023219
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,16384,12288,0.2754471037122938
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,6144,256,0.014541332920392355
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,512,64,0.0029066666546795103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,3584,32,0.005622222191757626
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,5120,64,0.007648888561460707
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,128,5120,0.009008888569143083
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,10240,12288,0.044827553961012095
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,2048,0.005345777918895085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,4096,2560,0.022826666633288067
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,128,1024,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,2560,0.008371555142932469
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,1536,1024,0.0063279999627007385
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2048,768,0.005598222215970357
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,256,6144,0.017629333668284945
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,32,1024,0.004000888930426704
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,16384,512,0.09208799733055963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,3072,3072,0.01624000072479248
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,1024,128,0.0035297779573334586
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,16384,0.11459378401438396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,2560,0.007672000262472365
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,10240,0.024816888901922438
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,1024,5120,0.009007111191749573
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,256,256,0.0053288886944452924
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,2048,32,0.0028640000770489373
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,512,7168,0.011158222125636207
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,1536,1536,0.006385777973466449
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,2048,2048,0.038902223110198975
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,2048,12288,0.13352178202735052
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,16384,3584,0.05638044410281711
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,65536,7168,0.2747804323832194
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,10240,16384,0.05625244643953112
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,128,7168,0.026722666290071275
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,16384,512,0.03184533450338576
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,4,64,16384,0.006649777707126405
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,8192,0.015999999311235216
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,3072,5120,0.011576000187132092
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,7168,1536,0.017850667238235474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,8192,1536,0.006719111154476802
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,256,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,5120,2560,0.01769244505299462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,768,256,0.0032631111227803757
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,256,65536,0.09808178080452813
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,2560,0.012878222597969903
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,128,65536,0.07946044206619263
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,3584,256,0.0058906666106647914
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,3584,1024,0.024474667178259954
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,65536,1536,0.10845777723524307
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,8192,4096,0.06082489093144735
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,12288,2048,0.011238222320874533
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,512,512,0.003636444608370463
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,6144,256,0.016525333126386006
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,10240,1024,0.007002666592597961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,3072,7168,0.01036888857682546
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,1536,10240,0.014228444960382251
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,128,0.0028400001012616684
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,256,6144,0.01073333372672399
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,16,1536,3072,0.006292444550328785
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,256,512,65536,0.06405688656700982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,64,8192,0.01129333343770769
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,3072,0.0199991116921107
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,5120,7168,0.02640266716480255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,768,2560,0.015487111277050443
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,256,1024,0.012075555821259817
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,4096,10240,0.017543110582563613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,768,65536,0.02583822276857164
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,16384,0.03500977820820279
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,3072,1024,0.006260444306664997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,7168,768,0.03283199999067519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,768,7168,0.020522667302025687
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,5120,65536,0.06977155473497179
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,4096,7168,0.05653866794374254
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,4096,4096,0.011772444678677453
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,5120,2560,0.0075786668393347
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,65536,3584,0.10624977615144517
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,4096,2560,0.018603555030292936
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8192,7168,128,0.03459022111362881
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8192,256,7168,0.05790755483839247
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,2048,3072,0.03399111164940728
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,256,65536,0.08088444338904487
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,8192,12288,0.03770844472779168
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,64,5120,0.008016000191370646
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,10240,12288,0.034697777695126004
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,3584,128,0.014749333262443542
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,7168,64,0.0029306666304667792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,128,128,0.00526311124364535
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,6144,128,0.014543111125628153
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,80,768,4096,0.016540444559521146
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,8192,512,0.007327111230956183
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,6144,1536,0.006583111153708563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,256,10240,0.012392000191741519
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,192,1536,2560,0.015784000356992085
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,768,6144,0.010876444478829702
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,7168,8192,0.032368001010682844
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2,3584,4096,0.020995555652512446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,10240,16384,0.11212889353434245
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2,128,10240,0.0069368887278768755
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,512,1024,0.00831911133395301
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,3072,8192,0.03328888946109348
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,8192,2560,0.02991022335158454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,160,256,5120,0.008010666403505538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,512,8192,0.02162222233083513
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,2560,64,0.005928888916969299
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,256,4096,0.015172445111804538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,3584,0.021189333664046392
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,2560,1536,0.007369777394665613
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,65536,6144,0.12434755431281196
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,512,65536,0.9870790905422635
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,7168,2560,0.007261333366235097
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,2048,5120,0.010189333723651038
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,6144,5120,0.023227555884255305
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,128,512,0.005569777968857024
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,12288,32,0.007707555260923173
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,768,128,0.0055440002017551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,8192,3072,0.01091733326514562
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,5120,8192,0.042938666211234204
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,768,128,0.0052382221652401825
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,256,6144,4096,0.026183999247021143
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,3072,12288,0.017825777331988018
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,768,512,0.003526222374704149
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,12288,1024,0.10188888841205174
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,65536,3072,3.409247080485026
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,3584,2048,0.015729778342776828
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,512,8192,0.027830223242441814
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,2560,1024,0.00582311095462905
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,2560,4096,0.014581332604090372
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,65536,3584,0.04487644301520454
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,64,128,1024,0.0038577777643998465
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,128,2048,0.009686222506894005
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,1536,256,0.006266666783226862
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,6144,65536,0.12965510951148138
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2,10240,3584,0.013753777576817406
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,32,2048,1536,0.004631999880075455
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,768,7168,0.018948444061809115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,2048,16384,0.04792622062895033
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,8,7168,256,0.005610666755172942
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,128,8192,0.044307556417253285
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,2048,5120,3072,0.04607644346025255
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,3584,256,0.005149333427349727
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,128,2048,0.015046222342385186
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,1024,3072,0.009396444592210982
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8192,3584,3584,0.1181173324584961
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1,256,3584,0.014568888478808932
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,512,1536,5120,0.011151110960377587
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,1,8192,6144,0.013957333233621387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,32,5120,0.009159111314349705
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,512,2560,0.007048889166778988
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,2560,65536,0.053211556540595166
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,2560,12288,0.03915111223856608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,768,7168,0.009895111123720804
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,128,12288,16384,0.04528888728883532
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,512,1024,4096,0.011953777737087674
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,256,65536,0.061398221386803516
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,512,256,16384,0.02244088881545597
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,16,5120,8192,0.01423466702302297
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,16,8192,10240,0.03255022234386868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,768,128,0.005190222213665645
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,1024,8192,0.01352355546421475
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,80,256,512,0.005936000082227919
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2048,256,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1,3584,65536,0.07390577925576104
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,1024,7168,0.02218311031659444
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,64,768,0.0038888889054457345
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,16,4096,1536,0.006695999867386288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,128,3584,0.006267555471923616
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,5120,1536,0.007983110845088959
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,2048,7168,256,0.009394666386975182
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,32,256,0.0028231110837724474
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,64,7168,1024,0.01720622181892395
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,256,0.005290666802061929
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,128,1024,0.005956444475385878
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,6144,6144,0.07609599828720093
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,16384,10240,0.0317795541551378
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,7168,7168,0.0384844442208608
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,2048,128,3072,0.012453333371215396
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,128,128,768,0.012744888663291931
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,2560,1536,0.006981333096822103
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,48,2560,16384,0.020649777518378366
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,32,10240,6144,0.030270222160551283
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,1536,7168,0.011774222056070963
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,128,128,0.002846222163902389
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4,2560,3584,0.007976888782448238
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,5120,768,0.004849777867396672
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,2560,4096,0.007673777639865875
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,768,12288,768,0.011741333537631564
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,3072,768,0.017622222503026325
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,7168,4096,0.04327466752794054
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,96,256,12288,0.008650666309727563
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,65536,2560,0.02777066661251916
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,256,2560,0.016728889611032274
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,2048,256,65536,0.17518756124708387
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,256,256,2560,0.005789333333571752
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,6144,1024,0.00628355559375551
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,48,512,6144,0.017839999662505258
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,256,65536,8192,0.07155999872419569
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,8192,2048,0.032560000816980995
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,80,65536,8192,0.17333689000871447
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,12288,6144,0.05323377913898892
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4,512,4096,0.015478221906556023
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,768,2048,0.016547555724779766
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,128,512,1024,0.0058986664646201665
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,48,16384,512,0.0052488889131281115
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,48,1024,12288,0.014263111684057446
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,64,2560,0.0059004442559348206
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,32,2560,7168,0.01273155543539259
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,768,10240,768,0.013407111167907715
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1024,512,65536,0.12627466519673666
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,32,3072,12288,0.01684888866212633
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,384,5120,4096,0.016653334101041157
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,192,512,4096,0.008632000121805403
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,64,128,768,0.005904888941182031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,96,5120,1536,0.006601777755551868
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,64,768,768,0.005591111050711737
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,8,6144,7168,0.02570399973127577
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8192,7168,5120,0.3260711034138997
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,384,3072,512,0.005959110955397288
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,160,65536,2048,0.03509155578083462
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,8,10240,1024,0.007320889168315464
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,96,5120,128,0.005599110904667113
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4,4096,65536,0.06656266583336724
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,1,1024,65536,0.05682488944795397
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,160,3072,2048,0.0069475554757648045
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,192,64,6144,0.006621333460013072
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,384,1536,128,0.014849777022997538
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,4096,8192,3584,0.10991733604007298
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,192,12288,128,0.006641777853171031
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8,4096,6144,16384,0.3202613459693061
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,1024,2560,7168,0.02985422147644891
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,768,8192,65536,0.4856773482428656
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,16384,768,0.04138488901986016
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,160,12288,128,0.016134222348531086
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,float16,128,65536,1536,0.04000622365209792
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,8,7168,1536,0.006287999865081575
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,96,7168,5120,0.026538666751649644
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,nvfp4,4096,65536,3072,0.2962515619066027
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,1024,512,512,0.01350311107105679
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,512,1024,5120,0.021407110823525324
SGLang,0.5.8.post1,NVIDIA GB200,gemm,sglang,fp8_block,768,2560,7168,0.035219556755489774
