framework,version,device,op_name,kernel_source,bmm_dtype,num_tokens,num_heads,latency
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,128,0.006703999638557434
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,16,0.004704000055789947
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,8,0.004633599892258644
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,4,0.004566400125622749
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,2,0.004492799937725067
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,1,0.004678399860858917
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,128,0.006937599927186966
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,64,0.004924799874424934
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,32,0.0047807998955249785
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,16,0.004502400010824204
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,8,0.004575999826192856
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,64,0.006332799792289734
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1,32,0.00456320010125637
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,4,0.004665600135922432
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,2,0.004793599992990494
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2,1,0.004460800066590309
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,128,0.006771200150251388
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,16,0.004579199850559235
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,64,0.006374400109052658
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,32,0.004467200115323066
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,8,0.0048096001148223875
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,2,0.0044895999133586885
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,4,0.004879999905824661
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4,1,0.004659200087189674
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,32,0.004611200094223023
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,128,0.006598400324583054
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,64,0.006278400123119354
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,16,0.0059424001723527905
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,4,0.0049727998673915865
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,2,0.004464000090956688
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,8,0.004659200087189674
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8,1,0.004812800139188766
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,64,0.005443200096487999
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,128,0.006572800129652024
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,32,0.004892800003290176
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,8,0.0046016000211238865
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,16,0.0046847999095916745
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,4,0.004873599857091904
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,1,0.0048096001148223875
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,16,2,0.004444799944758415
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,128,0.008755200356245042
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,32,0.004649600014090538
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,64,0.006547199934720993
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,16,0.0049727998673915865
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,8,0.004633599892258644
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,4,0.004950400069355964
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,2,0.004790399968624115
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,32,1,0.0047680001705884935
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,32,0.004720000177621841
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,64,0.006447999924421311
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,128,0.008582399785518646
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,16,0.004521600157022476
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,8,0.004931199923157692
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,4,0.0045056000351905824
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,2,0.004646399989724159
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,48,1,0.004899200052022934
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,128,0.006639999896287918
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,64,0.006319999694824219
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,32,0.006332799792289734
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,8,0.004473600164055824
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,16,0.004499199986457825
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,4,0.0047136001288890835
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,2,0.004483199864625931
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,64,1,0.004543999955058098
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,128,0.008508799970149994
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,32,0.006444799900054932
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,64,0.006473600119352341
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,16,0.006412799656391144
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,2,0.0050016000866889955
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,8,0.004742399975657463
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,4,0.0044511999934911724
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,80,1,0.004812800139188766
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,64,0.0065151996910572055
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,32,0.006729599833488464
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,16,0.004579199850559235
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,128,0.008703999966382981
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,8,0.005715199932456017
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,4,0.00432640016078949
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,1,0.004886399954557419
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,96,2,0.0046720001846551895
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,128,0.00854720026254654
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,64,0.006822399795055389
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,32,0.006742399930953979
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,16,0.005212799832224846
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,4,0.005529600009322166
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,8,0.00584000013768673
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,2,0.004732799902558327
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,128,1,0.004419200122356415
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,128,0.008966399729251862
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,32,0.006553599983453751
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,64,0.0066592000424861904
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,16,0.006329599767923355
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,4,0.004742399975657463
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,2,0.004534399881958961
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,1,0.004979199916124344
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,128,0.010585600137710571
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,32,0.0068000003695487974
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,64,0.008009599894285202
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,16,0.0065760001540184024
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,8,0.00517439991235733
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,2,0.004543999955058098
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,4,0.0044895999133586885
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,192,1,0.004793599992990494
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,128,0.010633599758148194
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,32,0.0065600000321865085
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,64,0.008646400272846222
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,16,0.006438399851322174
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,8,0.004598399996757508
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,4,0.004652800038456917
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,2,0.004720000177621841
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,256,1,0.0044064000248909
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,64,0.008579199761152267
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,128,0.012825599312782288
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,32,0.006623999774456024
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,16,0.006601600348949433
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,8,0.004860800132155419
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,2,0.004604800045490265
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,4,0.004646399989724159
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,320,1,0.004633599892258644
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,128,0.014758400619029999
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,64,0.010515200346708298
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,16,0.006496000289916992
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,32,0.008742400258779526
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,8,0.006534399837255478
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,4,0.00639680027961731
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,2,0.004560000076889992
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,160,8,0.004921599850058555
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,384,1,0.004681599885225296
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,128,0.018620799481868743
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,32,0.008550400286912918
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,64,0.012697599828243256
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,16,0.006595200300216675
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,4,0.004659200087189674
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,2,0.00461760014295578
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,8,0.006348799914121628
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,512,1,0.004793599992990494
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,128,0.02727360129356384
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,32,0.010729599744081497
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,64,0.014662399888038635
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,16,0.008432000130414962
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,8,0.006543999910354615
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,4,0.0065311998128890995
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,2,0.006319999694824219
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,768,1,0.004716800153255462
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,64,0.017766399681568144
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,32,0.010582400113344192
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,128,0.03541440069675446
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,16,0.00859839990735054
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,4,0.006492800265550614
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,8,0.006547199934720993
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,2,0.006115199998021126
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1024,1,0.004560000076889992
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,128,0.04758400022983551
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,64,0.02648960053920746
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,32,0.014828799664974213
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,16,0.010790400207042694
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,8,0.008486399799585343
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,4,0.00658240020275116
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,1,0.006297600269317627
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,1536,2,0.0066143997013568875
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,128,0.06294400095939637
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,64,0.034272000193595886
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,32,0.017004799842834473
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,8,0.009193599969148637
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,16,0.01080000028014183
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,4,0.006627199798822403
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,2,0.006563200056552887
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,2048,1,0.005369599908590317
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,128,0.08825280070304871
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,64,0.04668160080909729
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,16,0.014720000326633453
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,8,0.010601600259542465
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,32,0.026851201057434083
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,4,0.00865280032157898
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,2,0.006486400216817856
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,3072,1,0.006630399823188781
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,64,0.06242560148239136
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,32,0.03400320112705231
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,16,0.01690240055322647
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,128,0.11508799791336059
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,8,0.010790400207042694
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,4,0.008566399663686752
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,1,0.006307200342416763
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,4096,2,0.0065760001540184024
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,32,0.047891199588775635
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,64,0.08917760252952575
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,128,0.165174400806427
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,16,0.026399999856948853
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,8,0.014819200336933135
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,4,0.010601600259542465
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,2,0.008617600053548813
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,6144,1,0.006617599725723266
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,32,0.06487039923667907
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,64,0.11751999855041503
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,16,0.033456000685691836
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,128,0.22128639221191407
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,8,0.016867199540138246
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,4,0.010796800255775452
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,1,0.006460800021886826
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_pre,default,float16,8192,2,0.008640000224113464
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,128,0.006451199948787689
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,32,0.006499200314283371
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,16,0.006377600133419037
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,64,0.00628800019621849
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,8,0.006169600039720535
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,2,0.0049632001668214795
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,4,0.0056639999151229855
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1,1,0.0046016000211238865
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,64,0.006313599646091461
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,128,0.006451199948787689
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,32,0.006518399715423584
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,4,0.004681599885225296
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,8,0.006275200098752975
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,16,0.00626240000128746
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,2,0.00459199994802475
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2,1,0.004732799902558327
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,128,0.0063680000603199005
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,64,0.005324799939990044
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,32,0.00634239986538887
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,16,0.006400000303983688
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,8,0.006489600241184235
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,4,0.004931199923157692
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,2,0.004822399839758873
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4,1,0.004595199972391129
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,128,0.006255999952554703
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,64,0.006092799827456474
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,32,0.006425599753856659
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,8,0.006268800050020218
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,4,0.0046847999095916745
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,16,0.0062431998550891874
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,2,0.006169600039720535
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8,1,0.004662400111556053
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,128,0.006566400080919266
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,64,0.006310400366783142
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,32,0.005263999849557876
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,8,0.006239999830722809
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,4,0.004710400104522705
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,16,0.006255999952554703
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,2,0.006185600161552429
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16,1,0.004694399982690811
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,128,0.006598400324583054
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,32,0.0063008002936840056
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,64,0.006284800171852112
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,8,0.005366399884223938
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,16,0.006227200105786324
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,4,0.006227200105786324
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,2,0.0044096000492572784
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,32,1,0.006089600175619126
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,128,0.006758400052785873
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,64,0.006310400366783142
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,32,0.006425599753856659
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,16,0.00628800019621849
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,8,0.005075199902057648
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,4,0.006275200098752975
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,2,0.006060799956321717
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,48,1,0.0046431999653577805
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,128,0.006790400296449661
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,64,0.006259199976921081
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,32,0.006345599889755249
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,16,0.00623680017888546
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,4,0.004646399989724159
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,8,0.006080000102519989
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,1,0.00448639988899231
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,64,2,0.00639680027961731
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,128,0.006752000004053116
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,32,0.006246399879455566
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,64,0.006550399959087372
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,16,0.006185600161552429
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,8,0.0063231997191905975
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,4,0.004704000055789947
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,2,0.006224000081419945
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,80,1,0.0060479998588562015
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,128,0.00647680014371872
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,32,0.006371200084686279
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,16,0.005116799846291542
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,64,0.006246399879455566
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,8,0.00634239986538887
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,4,0.0047391999512910845
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,2,0.005628800019621849
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,96,1,0.0050432000309228895
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,32,0.007846400141716003
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,64,0.006387200206518173
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,128,0.006732799857854843
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,16,0.006255999952554703
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,4,0.004668800160288811
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,8,0.006304000318050384
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,1,0.004492799937725067
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,128,2,0.005132799968123436
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,128,0.008499199897050858
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,32,0.006387200206518173
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,64,0.006492800265550614
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,16,0.006278400123119354
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,8,0.006348799914121628
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,2,0.005638400092720985
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,4,0.005023999884724617
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,160,1,0.0051807999610900875
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,64,0.006496000289916992
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,128,0.008390399813652038
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,32,0.0063231997191905975
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,16,0.006447999924421311
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,4,0.006390400230884552
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,8,0.006191999837756157
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,2,0.0061471998691558834
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,192,1,0.006214400008320809
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,128,0.008966399729251862
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,64,0.006684800237417221
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,32,0.006304000318050384
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,16,0.006428799778223038
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,8,0.006281600147485733
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,4,0.00623680017888546
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,2,0.006284800171852112
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,256,1,0.006025600060820579
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,128,0.01050880029797554
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,64,0.00828159973025322
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,16,0.006451199948787689
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,32,0.006812799721956253
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,4,0.006431999802589417
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,8,0.006403200328350067
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,2,0.005795200169086456
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,320,1,0.005407999828457832
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,128,0.01053439974784851
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,64,0.00846719965338707
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,32,0.006716799736022949
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,8,0.006143999844789505
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,16,0.006467200070619583
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,4,0.006310400366783142
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,1,0.005734400078654289
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,384,2,0.006297600269317627
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,64,0.00859839990735054
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,128,0.015065599977970124
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,8,0.006294400244951248
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,16,0.006454399973154068
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,4,0.0062720000743865965
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,2,0.006060799956321717
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,1,0.006198399886488915
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,128,0.029017600417137145
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,64,0.010467199981212616
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,16,0.006764800101518631
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,32,0.008441600203514098
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,4,0.006364800035953522
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,8,0.006204799935221672
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,2,0.006275200098752975
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,768,1,0.006143999844789505
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,128,0.03313280045986176
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,64,0.01252799928188324
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,16,0.007065600156784058
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,32,0.008799999952316284
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,8,0.006393600255250931
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,4,0.006393600255250931
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,2,0.006291200220584869
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1024,1,0.00618240013718605
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,64,0.025097599625587462
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,128,0.04778240025043488
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,32,0.010652799904346467
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,16,0.008588799834251403
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,8,0.006918399780988693
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,2,0.006380800157785416
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,4,0.006224000081419945
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,1536,1,0.006412799656391144
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,64,0.03338879942893982
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,128,0.06334400177001953
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,32,0.012646399438381195
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,16,0.008752000331878663
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,8,0.006809599697589874
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,4,0.006332799792289734
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,2,0.006291200220584869
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,2048,1,0.006291200220584869
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,64,0.0460671991109848
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,128,0.08534079790115356
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,32,0.026851201057434083
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,16,0.010598400235176086
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,8,0.008604799956083297
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,4,0.006700800359249115
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,2,0.006246399879455566
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,3072,1,0.006294400244951248
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,128,0.11945279836654663
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,64,0.06267840266227723
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,16,0.01268479973077774
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,512,32,0.006758400052785873
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,32,0.03363839983940124
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,8,0.008767999708652496
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,4,0.006854400038719177
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,1,0.0069983996450901035
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,4096,2,0.006329599767923355
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,64,0.08655679821968079
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,128,0.15836479663848876
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,32,0.04663360118865967
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,16,0.024857600033283234
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,8,0.01040000021457672
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,4,0.008502399921417237
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,2,0.006780800223350525
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,6144,1,0.006483200192451477
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,32,0.06346880197525025
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,64,0.12113280296325683
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,16,0.031196799874305726
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,128,0.2234015941619873
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,8,0.012918399274349212
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,4,0.008780799806118011
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,1,0.00634239986538887
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,8192,2,0.006572800129652024
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,64,0.17224960327148436
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,128,0.3325952053070068
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,32,0.0896992027759552
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,8,0.023977600038051605
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,16,0.045491200685501096
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,4,0.010396800190210342
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,2,0.008454400300979614
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,12288,1,0.006579200178384781
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,64,0.22966721057891845
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,128,0.44667840003967285
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,32,0.1286463975906372
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,8,0.03136000037193298
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,16,0.06142079830169678
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,4,0.01340160071849823
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,1,0.006844799965620041
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,16384,2,0.008851200342178345
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,64,0.28260159492492676
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,32,0.15204800367355348
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,128,0.5564479827880859
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,16,0.07487040162086486
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,4,0.01686079949140549
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,8,0.03950079977512359
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,2,0.008787199854850769
TRTLLM,1.1.0,NVIDIA GB200,mla_gen_post,default,float16,20480,1,0.008326400071382523
