framework,version,device,op_name,kernel_source,bmm_dtype,num_tokens,num_heads,latency
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,128,0.008665599673986436
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,64,0.006345599889755249
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,32,0.005811199918389321
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,128,0.01273919939994812
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,16,0.004633599892258644
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,16,0.009804800152778625
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,8,0.004806400090456009
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,8,0.010604800283908844
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,4,0.006406400352716446
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,4,0.0106175996363163
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,2,0.006294400244951248
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,2,0.01003199964761734
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1,1,0.004518400132656098
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,1,0.00862400010228157
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,128,0.008528000116348267
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,64,0.008899199962615966
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,64,0.006364800035953522
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1,32,0.008761599659919739
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,32,0.005324799939990044
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,128,0.012915199995040894
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,16,0.005395200103521347
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,64,0.010684800148010255
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,16,0.0107744000852108
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,8,0.008739200234413148
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,8,0.006224000081419945
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,4,0.004710400104522705
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,2,0.004675199836492538
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,4,0.010598400235176086
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,2,0.008953599631786347
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2,1,0.006118400022387505
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,128,0.008710400015115739
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,1,0.00888959988951683
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,64,0.00634239986538887
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,128,0.01353600025177002
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,32,0.006275200098752975
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,64,0.010598400235176086
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,16,0.005142400041222572
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,32,0.010598400235176086
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,16,0.010572800040245056
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,8,0.004966399818658829
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,8,0.01053759977221489
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,4,0.006035200133919716
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,2,0.0065151996910572055
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4,1,0.004745600000023842
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,2,0.008796799927949905
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,128,0.0086496002972126
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,1,0.008729600161314011
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,64,0.006598400324583054
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,128,0.014604799449443817
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,32,0.0061055999249219894
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,64,0.010585600137710571
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,16,0.004659200087189674
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,32,0.010576000064611435
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,8,0.0047775998711586
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,16,0.008627200126647949
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,4,0.0046847999095916745
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,8,0.008921600133180618
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,4,0.008739200234413148
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,2,0.005955199897289276
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8,1,0.004889599978923798
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,2,0.010751999914646149
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8,1,0.00883520022034645
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,128,0.008627200126647949
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,64,0.006592000275850296
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2,32,0.01072319969534874
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,32,0.006143999844789505
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,64,0.010777600109577179
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,128,0.01496960073709488
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,16,0.00488319993019104
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,16,0.010601600259542465
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,8,0.004694399982690811
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,8,0.01051200032234192
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,4,0.004665600135922432
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,2,0.004681599885225296
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,4,0.010582400113344192
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,2,0.010710400342941285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,16,1,0.0060095999389886854
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,1,0.010716799646615982
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,128,0.008848000317811966
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,64,0.006496000289916992
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,128,0.018822400271892546
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,32,0.006335999816656113
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,64,0.012620800733566284
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,32,0.010915199667215348
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,16,0.006406400352716446
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,16,32,0.010857599973678588
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,8,0.0048096001148223875
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,16,0.010585600137710571
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,8,0.010761599987745285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,4,0.004761600121855736
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,4,0.008771199733018875
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,2,0.0064191997051239015
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,2,0.008918400108814239
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,32,1,0.004735999926924706
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,128,0.008950400352478027
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,32,1,0.01072319969534874
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,128,0.020972800254821778
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,64,0.006441599875688553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,32,0.006438399851322174
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,64,0.013081599771976472
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,32,0.010947199910879135
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,16,0.005529600009322166
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,8,0.004927999898791313
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,16,0.010950399935245514
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,4,0.006195199862122536
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,8,0.010681600123643876
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,2,0.005427199974656105
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,4,0.010553599894046783
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,48,1,0.0048191998153924945
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,2,0.010608000308275222
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,48,1,0.008659200370311737
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,128,0.008560000360012055
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,64,0.0065151996910572055
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,128,0.023081600666046143
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,64,0.01478399932384491
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,32,0.006406400352716446
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,16,0.006438399851322174
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,32,0.012646399438381195
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,16,0.010726399719715118
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,8,0.004854400083422661
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,4,0.004944000020623207
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,8,0.010806400328874588
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,2,0.006092799827456474
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,4,0.00984639972448349
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,2,0.010569600015878677
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,64,1,0.004502400010824204
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,64,1,0.008736000210046769
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,128,0.008588799834251403
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,64,0.006489600241184235
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,128,0.02526719868183136
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,64,0.01698880046606064
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,32,0.006451199948787689
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,16,0.006035200133919716
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,32,0.01276479959487915
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,8,0.004499199986457825
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,16,0.01090560033917427
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,4,0.006259199976921081
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,8,0.01064319983124733
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,2,0.004691199958324432
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,4,0.010659199953079224
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,80,1,0.006121600046753883
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,2,0.008665599673986436
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,80,1,0.010566399991512298
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,128,0.008611200004816055
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,64,0.006518399715423584
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,128,0.029177600145339967
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,32,0.006611199676990509
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,64,0.018943999707698823
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,32,0.012918399274349212
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,16,0.006415999680757523
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,8,0.006329599767923355
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,16,0.010784000158309937
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,4,0.0061760000884532925
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,8,0.01061440035700798
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,4,0.010585600137710571
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,2,0.0064351998269557955
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,96,1,0.005606399849057198
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,2,0.010387200117111205
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,96,1,0.010627199709415436
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,128,0.008720000088214875
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,128,0.03313600122928619
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,64,0.006790400296449661
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,32,0.006492800265550614
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,64,0.02091200053691864
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,16,0.006284800171852112
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,32,0.01496639996767044
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,8,0.006294400244951248
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,16,0.011385600268840789
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,4,0.004515200108289719
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,8,0.010467199981212616
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,2,0.005219199880957604
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,4,0.010531199723482132
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,128,1,0.004867200180888176
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,2,0.010611200332641601
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,128,1,0.010633599758148194
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,128,0.010527999699115753
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,64,0.008451200276613235
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,128,0.039324799180030824
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,32,0.006592000275850296
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,64,0.024835200607776643
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,16,0.006441599875688553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,32,0.016710400581359863
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,16,0.01297920048236847
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,8,0.006521599739789963
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,4,0.0063680000603199005
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,8,0.012223999947309494
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,4,0.010755199939012527
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,2,0.006185600161552429
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,160,1,0.0047775998711586
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,2,0.010585600137710571
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,160,1,0.00878399983048439
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,128,0.010710400342941285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,64,0.008627200126647949
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,128,0.04339520037174225
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,64,0.027024000883102417
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,32,0.006579200178384781
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,16,0.006521599739789963
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,32,0.016873599588871004
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,8,0.00647680014371872
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,16,0.012825599312782288
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,8,0.010780800133943558
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,4,0.00639680027961731
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,4,0.01072319969534874
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,2,0.006403200328350067
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,2,0.01064639985561371
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,192,1,0.004582399874925614
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,192,1,0.008601599931716919
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,128,0.012656000256538392
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,128,0.055743998289108275
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,64,0.0086496002972126
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,64,0.03312000036239624
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,32,0.0067103996872901915
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,32,0.020873600244522096
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,16,0.006537599861621857
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,8,0.006233600154519081
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,8,0.011049599945545196
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,16,0.015305599570274353
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,4,0.0048191998153924945
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,2,0.004499199986457825
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4,4,0.008819200098514557
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,4,0.01170239970088005
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,2,0.010566399991512298
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,256,1,0.01074879989027977
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,128,0.014608000218868256
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,64,0.01061440035700798
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,128,0.06609920263290406
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,32,0.007983999699354172
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,64,0.03939839899539947
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,16,0.006505600363016129
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,32,0.023081600666046143
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,16,0.017132799327373504
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,8,0.006563200056552887
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,4,0.006220800057053566
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,8,0.012835200130939483
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,2,0.004745600000023842
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,4,0.01080000028014183
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,2,0.010639999806880952
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,320,1,0.008723200112581254
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,320,1,0.006291200220584869
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,128,0.015119999647140503
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,128,0.07854080200195312
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,64,0.010649599879980088
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,32,0.008556800335645676
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,64,0.04364160001277924
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,256,1,0.004566400125622749
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,32,0.027024000883102417
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,16,0.006681600213050842
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,8,0.00658240020275116
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,16,0.017683200538158417
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,8,0.014636799693107605
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,4,0.0065600000321865085
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,2,0.01064319983124733
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,2,0.006496000289916992
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,4,0.012742400169372559
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,384,1,0.006191999837756157
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,384,1,0.010361599922180175
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,128,0.019142399728298187
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,128,0.10299520492553711
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,64,0.012800000607967377
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,64,0.05609920024871826
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,32,0.008575999736785888
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,32,0.032630398869514465
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,16,0.006684800237417221
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,8,0.006483200192451477
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,16,0.020870399475097657
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,8,0.014767999947071075
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,4,0.006291200220584869
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,4,0.010700800269842149
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,1,0.0049183998256921765
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,512,2,0.006358399987220764
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,2,0.010428799688816071
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,512,1,0.0106175996363163
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,64,0.015305599570274353
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,128,0.027049601078033447
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,32,0.010732799768447876
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,128,0.15143680572509766
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,64,0.07828800082206726
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,16,0.008825600147247314
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,32,0.04542720019817352
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,8,0.006723199784755707
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,16,0.026848000288009644
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,8,0.018694399297237395
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,4,0.006870400160551071
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,4,0.012806400656700134
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,2,0.006294400244951248
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,768,1,0.005126399919390678
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,2,0.012646399438381195
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,768,1,0.010700800269842149
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,128,0.03489919900894165
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,64,0.01881600022315979
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,32,0.012604799866676331
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,64,0.10169600248336792
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,128,0.1946079969406128
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,16,0.008553600311279297
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,32,0.05577279925346375
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,8,0.006550399959087372
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,16,0.033206400275230405
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,4,0.006569600105285645
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,8,0.02099519968032837
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,2,0.006252799928188324
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,4,0.015145599842071533
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1024,1,0.0063680000603199005
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,2,0.010751999914646149
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1024,1,0.010886400192975997
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,128,0.04758720099925995
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,64,0.02698560059070587
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,128,0.2937567949295044
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,64,0.14882559776306153
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,32,0.015689599514007568
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,16,0.012579199671745301
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,32,0.07816320061683654
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,8,0.008617600053548813
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,16,0.045475199818611145
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,4,0.0065760001540184024
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,4,0.01849920004606247
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,8,0.02720319926738739
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,2,0.0065311998128890995
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,2,0.012777599692344665
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,1536,1,0.0063680000603199005
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,1536,1,0.012643200159072877
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,128,0.060012799501419065
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,64,0.03386879861354828
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,32,0.018806399405002595
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,64,0.19701759815216063
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,128,0.3842144012451172
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,16,0.012534399330615998
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,32,0.10342719554901122
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,8,0.008787199854850769
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,16,0.05568000078201294
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,4,0.0077855996787548065
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,8,0.033190399408340454
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,4,0.021081599593162536
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,2,0.006489600241184235
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,2048,1,0.006409599632024765
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,2,0.014710399508476257
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,2048,1,0.012534399330615998
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,128,0.08552640080451965
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,64,0.04731839895248413
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,32,0.02714880108833313
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,128,0.566099214553833
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,64,0.29405760765075684
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,32,0.15186879634857178
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,16,0.014876799285411834
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,8,0.012595200538635254
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,16,0.07822719812393189
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,4,0.00873280018568039
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,2,0.006815999746322632
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,4,0.027030399441719054
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,8,0.04542720019817352
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,2,0.018726399540901183
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,3072,1,0.006691200286149978
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,3072,1,0.014739200472831726
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,128,0.11020799875259399
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,64,0.059494400024414064
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,32,0.03340800106525421
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,64,0.3912928104400635
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,32,0.1931679964065552
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,128,0.7497024059295654
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,16,0.01863040030002594
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,16,0.1028480052947998
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,8,0.012630400061607362
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,8,0.05782080292701721
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,4,0.008617600053548813
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,2,0.007811199873685837
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,4,0.032425600290298465
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,4096,1,0.006329599767923355
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,2,0.021507200598716737
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,4096,1,0.016089600324630738
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,128,0.15959999561309815
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,64,0.08514879941940308
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,32,0.04795520007610321
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,32,0.29347519874572753
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,64,0.5768032073974609
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,16,0.02746880054473877
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,16,0.15051840543746947
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,8,0.014975999295711518
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,128,1.1214624404907227
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,8,0.07840319871902465
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,4,0.011347199976444244
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,4,0.04544320106506348
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,2,0.008793599903583527
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,6144,1,0.006524799764156342
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,2,0.02693440020084381
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,6144,1,0.019068799912929535
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,64,0.11001919507980347
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,128,0.20680320262908936
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,32,0.05984640121459961
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,32,0.38273599147796633
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,64,0.7647647857666016
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,16,0.03374080061912536
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,8,0.01706240028142929
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,16,0.1969696044921875
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,128,1.4987168312072754
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,8,0.10279359817504882
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,4,0.012479999661445617
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,2,0.008643200248479843
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,4,0.05584959983825684
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,2,0.033257600665092465
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,float16,8192,1,0.006521599739789963
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_pre,default,fp8,8192,1,0.02242880016565323
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,128,0.006335999816656113
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,128,0.008841600269079208
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,64,0.00623680017888546
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,64,0.010812799632549285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,32,0.006550399959087372
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,32,0.010662399977445603
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,16,0.006700800359249115
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,16,0.010809600353240967
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,8,0.006729599833488464
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,8,0.009734400361776353
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,4,0.006220800057053566
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,4,0.008716800063848496
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,2,0.006294400244951248
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,2,0.010623999685049058
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1,1,0.006473600119352341
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1,1,0.010684800148010255
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,128,0.006534399837255478
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,128,0.010592000186443329
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,64,0.0064640000462532045
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,64,0.009945599734783173
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,32,0.006441599875688553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,32,0.008726400136947633
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,16,0.0064191997051239015
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,16,0.010819199681282043
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,8,0.006592000275850296
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,8,0.010771200060844421
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,4,0.006332799792289734
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,4,0.008790399879217148
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,2,0.006380800157785416
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,2,0.008668799698352814
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2,1,0.004732799902558327
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,128,0.006489600241184235
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,128,0.010707200318574906
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,64,0.006281600147485733
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,64,0.010710400342941285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,32,0.0065151996910572055
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,32,0.010630399733781815
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,16,0.006585600227117539
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,16,0.010659199953079224
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,8,0.006550399959087372
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,8,0.008908800035715102
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,4,0.005779200047254562
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,4,0.010569600015878677
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,2,0.00623680017888546
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,2,0.008633600175380706
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4,1,0.006307200342416763
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4,1,0.008591999858617782
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,128,0.0065760001540184024
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,128,0.010678400099277497
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,64,0.006447999924421311
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,64,0.010665600001811982
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,32,0.006345599889755249
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,32,0.008867199718952178
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,16,0.006566400080919266
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,16,0.009036800265312195
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,8,0.00639680027961731
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,8,0.010710400342941285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,4,0.006454399973154068
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,4,0.010688000172376633
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,2,0.006313599646091461
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,2,0.010566399991512298
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8,1,0.007084800302982331
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8,1,0.008566399663686752
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,128,0.006579200178384781
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,128,0.012636800110340119
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,64,0.006441599875688553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,64,0.010710400342941285
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,32,0.006281600147485733
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,32,0.010694400221109391
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,16,0.006220800057053566
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,16,0.00867839977145195
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,8,0.006579200178384781
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,8,0.010598400235176086
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,4,0.006185600161552429
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,4,0.01051200032234192
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,2,0.010681600123643876
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,1,0.006470400094985962
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16,1,0.010627199709415436
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,128,0.00660799965262413
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,128,0.01515199989080429
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,64,0.006415999680757523
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,64,0.011766400188207626
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,32,0.006540799885988236
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,32,0.01072319969534874
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,16,0.006406400352716446
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,16,0.008947200328111648
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,8,0.006073600053787232
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,8,0.010704000294208527
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,4,0.00642239972949028
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,4,0.010595200210809707
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,2,0.006095999851822853
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,2,0.008591999858617782
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,32,1,0.004694399982690811
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,32,1,0.009539200365543366
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,128,0.00655680000782013
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,128,0.01880960017442703
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,64,0.006553599983453751
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,64,0.014796799421310425
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,32,0.006425599753856659
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,32,0.010755199939012527
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16,2,0.0062720000743865965
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,16,0.006380800157785416
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2,1,0.00875839963555336
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,8,0.006377600133419037
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,8,0.008899199962615966
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,4,0.006361600011587143
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,4,0.010675200074911118
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,2,0.006412799656391144
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,2,0.01053439974784851
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,48,1,0.006457599997520447
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,1,0.010569600015878677
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,128,0.006764800101518631
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,128,0.020729599893093108
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,64,0.006569600105285645
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,64,0.014912000298500061
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,32,0.00647680014371872
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,32,0.01268800050020218
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,16,0.006579200178384781
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,16,0.010540799796581268
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,8,0.0063008002936840056
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,8,0.010467199981212616
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,4,0.0064351998269557955
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,4,0.010476800054311753
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,2,0.00642239972949028
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,2,0.010582400113344192
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,64,1,0.006441599875688553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,64,1,0.010745599865913391
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,128,0.008457600325345992
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,128,0.024326400458812715
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,64,0.006592000275850296
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,64,0.016681599617004394
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,32,0.006598400324583054
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,32,0.012796799838542938
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,16,0.006393600255250931
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,16,0.010864000022411346
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,8,0.006387200206518173
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,8,0.010451199859380722
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,4,0.006534399837255478
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,4,0.01085119992494583
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,2,0.0063680000603199005
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,2,0.010611200332641601
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,80,1,0.005584000051021576
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,80,1,0.009478399902582169
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,128,0.007132799923419952
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,128,0.02648960053920746
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,64,0.006585600227117539
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,64,0.01873279958963394
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,32,0.006665600091218948
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,32,0.014633600413799287
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,16,0.012598399817943574
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,16,0.006831999868154526
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,8,0.006326399743556976
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,4,0.006537599861621857
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,8,0.011286400258541107
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,4,0.010524799674749374
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,2,0.007014399766921997
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,96,1,0.006409599632024765
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,2,0.010460799932479859
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,96,1,0.008646400272846222
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,128,0.008352000266313553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,64,0.006508799642324448
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,128,0.031017601490020752
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,64,0.020716799795627593
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,32,0.006307200342416763
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,16,0.006393600255250931
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,32,0.014713600277900696
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,16,0.01064639985561371
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,8,0.006483200192451477
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,48,16,0.010611200332641601
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,4,0.006371200084686279
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,8,0.01055999994277954
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,2,0.00639680027961731
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,128,1,0.006377600133419037
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,4,0.010691200196743012
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,2,0.010652799904346467
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,128,1,0.010598400235176086
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,128,0.008428800106048583
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,128,0.0372512012720108
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,64,0.006857600063085556
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,32,0.006387200206518173
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,64,0.022976000607013703
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,16,0.006294400244951248
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,32,0.017110399901866913
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,8,0.006377600133419037
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,16,0.012886400520801543
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,8,0.010886400192975997
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,4,0.006358399987220764
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,2,0.006166400015354156
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,4,0.010835199803113937
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,2,0.010505600273609162
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,160,1,0.0063680000603199005
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,160,1,0.010428799688816071
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,128,0.008524800091981888
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,128,0.041340801119804385
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,64,0.006838399916887283
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,64,0.025158399343490602
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,32,0.006668800115585327
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,16,0.006470400094985962
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,32,0.01706559956073761
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,8,0.006550399959087372
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,16,0.01456640064716339
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,8,0.012521600723266602
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,4,0.006265600025653839
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,4,0.010819199681282043
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,2,0.00618240013718605
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,192,1,0.006403200328350067
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,2,0.010505600273609162
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,192,1,0.00997759997844696
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,128,0.008934400230646133
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,64,0.00814720019698143
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,128,0.051849597692489625
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,64,0.030972799658775328
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,32,0.006335999816656113
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,32,0.02025599926710129
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,16,0.006387200206518173
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,8,0.006579200178384781
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,16,0.014934399724006652
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,4,0.006332799792289734
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,8,0.012588800489902496
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,2,0.0064640000462532045
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,4,0.010822399705648422
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,2,0.01085119992494583
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,256,1,0.006438399851322174
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,256,1,0.010576000064611435
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,128,0.010377600044012069
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,64,0.008620800077915191
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,128,0.0639136016368866
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,32,0.006646399945020675
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,64,0.03740800023078918
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,16,0.006579200178384781
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,32,0.02306240051984787
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,8,0.006569600105285645
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,16,0.01666239947080612
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,4,0.0063391998410224915
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,8,0.012720000743865967
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,2,0.006467200070619583
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,2,0.010422399640083313
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,320,1,0.006073600053787232
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,1,0.010444799810647965
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,128,0.010883200168609618
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,64,0.008633600175380706
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,128,0.07613120079040528
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,32,0.006700800359249115
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,64,0.041388800740242
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,16,0.006374400109052658
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,32,0.026844799518585205
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,8,0.0064640000462532045
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,16,0.017712000012397765
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,4,0.006307200342416763
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,8,0.012912000715732574
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,4,0.012556800246238708
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,2,0.006409599632024765
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,384,1,0.00628800019621849
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,2,0.010780800133943558
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,384,1,0.010572800040245056
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,128,0.015244799852371215
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,64,0.008928000181913375
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,128,0.10241600275039672
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,64,0.05185279846191406
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,32,0.00835840031504631
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,16,0.0066592000424861904
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,32,0.03121280074119568
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,8,0.006752000004053116
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,16,0.02120320051908493
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,4,0.006332799792289734
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,8,0.015056000649929046
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,4,0.01265919953584671
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,2,0.006441599875688553
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,2,0.01069760024547577
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,512,1,0.006595200300216675
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,512,1,0.010454399883747101
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,128,0.02898559868335724
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,64,0.010918399691581726
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,128,0.15012480020523072
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,32,0.008844800293445587
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,64,0.07624959945678711
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,16,0.006745599955320358
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,32,0.041510400176048276
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,16,0.025222399830818178
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,8,0.006454399973154068
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,8,0.016982400417327882
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,4,0.006371200084686279
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,2,0.006483200192451477
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,4,0.012873600423336028
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,2,0.01096320003271103
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,768,1,0.006364800035953522
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,768,1,0.010550399869680404
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,128,0.03636159896850586
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,128,0.19552639722824097
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,64,0.012892800569534301
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,32,0.009232000261545182
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,64,0.10086079835891723
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,32,0.05317760109901428
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,16,0.0084927998483181
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,8,0.006617599725723266
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,16,0.031411200761795044
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,8,0.020777599513530733
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,4,0.006499200314283371
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,4,0.014585599303245544
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,2,0.006268800050020218
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,2,0.012652799487113953
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1024,1,0.006294400244951248
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1024,1,0.010896000266075134
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,128,0.0478879988193512
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,64,0.027641600370407103
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,128,0.2895776033401489
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,32,0.010780800133943558
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,32,0.07573440074920654
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,64,0.14951679706573487
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,16,0.008294399827718735
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,8,0.006524799764156342
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,8,0.025164800882339477
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,16,0.04239999949932098
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,4,0.006384000182151794
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,2,0.00650240033864975
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,4,0.017215999960899352
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,2,0.012748800218105316
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,1536,1,0.006649599969387054
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,1536,1,0.0118367999792099
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,128,0.06436799764633179
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,128,0.38251519203186035
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,64,0.03363839983940124
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,32,0.012649600207805634
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,64,0.19498239755630492
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,16,0.009257599711418152
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,32,0.10035840272903443
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,8,0.006812799721956253
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,16,0.05340480208396912
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,4,0.0065600000321865085
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,8,0.031017601490020752
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,4,0.0208064004778862
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,2,0.006543999910354615
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,2048,1,0.006265600025653839
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,2,0.014659200608730317
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,2048,1,0.012697599828243256
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,128,0.08661440014839172
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,64,0.046009600162506104
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,64,0.2871999979019165
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,128,0.5810688018798829
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,32,0.027036800980567932
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,16,0.010502400249242783
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,32,0.14967360496520996
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,16,0.07455679774284363
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,8,0.00851840004324913
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,8,0.041331198811531064
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,4,0.006527999788522721
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,4,0.024940800666809083
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,2,0.006492800265550614
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,3072,1,0.006505600363016129
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,2,0.01706240028142929
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,3072,1,0.012774400413036346
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,128,0.1132383942604065
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,64,0.05994880199432373
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,64,0.3822751998901367
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,128,0.7945119857788085
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,32,0.031196799874305726
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,16,0.012675200402736665
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,32,0.195632004737854
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,16,0.10239039659500122
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,8,0.008703999966382981
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,4,0.006611199676990509
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,8,0.05351679921150208
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,2,0.0063391998410224915
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,4,0.03096640110015869
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,4096,1,0.006220800057053566
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,2,0.020819200575351714
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,4096,1,0.014732800424098969
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,128,0.16392960548400878
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,64,0.08285120129585266
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,64,0.573692798614502
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,32,0.04514240026473999
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,128,1.2477919578552246
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,16,0.025385600328445435
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,32,0.2856287956237793
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,8,0.0107744000852108
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,8,0.07618560194969178
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,16,0.15
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,4,0.008713600039482117
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,4,0.04171839952468872
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,2,0.006764800101518631
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,6144,1,0.0066143997013568875
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,2,0.025862398743629455
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,6144,1,0.01701440066099167
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,128,0.21157441139221192
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,64,0.10778239965438843
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,32,0.060726398229599
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,128,1.724505615234375
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,64,0.7981823921203614
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,320,4,0.01090560033917427
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,32,0.3813024044036865
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,16,0.03268480002880096
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,16,0.19509119987487794
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,4,0.00963200032711029
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,2,0.008345600217580795
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,8,0.10069119930267334
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,4,0.0515936017036438
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,1,0.006841599941253662
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,2,0.03139840066432953
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,8192,1,0.019305600225925444
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,64,0.15828800201416016
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,128,0.3111104011535645
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,64,1.226249599456787
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,8192,8,0.012870399653911591
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,32,0.08599680066108703
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,16,0.045228800177574156
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,128,2.642086410522461
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,16,0.2867808103561401
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,8,0.027449598908424376
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,32,0.5693888187408447
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,4,0.010531199723482132
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,2,0.008607999980449676
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,8,0.14777599573135375
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,4,0.0762943983078003
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,12288,1,0.006950400024652481
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,2,0.041500800848007204
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,12288,1,0.025411200523376466
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,64,0.20561599731445312
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,128,0.4090432167053223
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,32,0.10688320398330689
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,32,0.7836448192596436
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,16,0.0597536027431488
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,64,1.7004159927368163
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,16,0.3812031984329224
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,8,0.03128640055656433
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,4,0.012774400413036346
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,8,0.1951807975769043
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,2,0.010115200281143188
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,2,0.053439998626708986
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,4,0.10091199874877929
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,16384,1,0.008563199639320373
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,1,0.031033599376678468
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,16384,128,3.5743328094482423
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,64,0.2519551992416382
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,128,0.5095392227172851
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,32,0.13037760257720948
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,32,1.0096832275390626
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,16,0.07104640007019043
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,64,2.1642528533935548
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,8,0.039299198985099794
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,16,0.47531518936157224
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,4,0.017238399386405943
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,8,0.23845438957214354
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,4,0.1249567985534668
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,2,0.010444799810647965
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,float16,20480,1,0.008489599823951722
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,2,0.06399040222167969
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,1,0.037206399440765384
SGLang,0.5.8.post1,NVIDIA GB200,mla_gen_post,default,fp8,20480,128,4.479308700561523
