framework,version,device,op_name,kernel_source,bmm_dtype,num_tokens,num_heads,latency
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,128,0.006278400123119354
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,64,0.005049600079655647
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,32,0.006467200070619583
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,128,0.015868799388408662
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,64,0.010742399841547012
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,8,0.007062400132417679
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,4,0.0041184000670909885
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,16,0.004291199892759323
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,8,0.006124800071120262
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,2,0.0040640000253915785
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1,1,0.004198399931192398
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,32,0.008249600231647492
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,4,0.006403200328350067
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,128,0.0064640000462532045
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,16,0.007248000055551529
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,2,0.0060639999806880954
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,64,0.005014400184154511
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,64,0.010761599987745285
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,32,0.004572800174355507
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1,1,0.005833600088953972
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,128,0.01621119976043701
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,16,0.004287999868392944
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,8,0.006918399780988693
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,32,0.008460800349712371
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,4,0.0038943998515605927
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,8,0.004447999969124794
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,2,0.004204799979925155
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,16,0.00729919970035553
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,1,0.006224000081419945
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,128,0.016361600160598753
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,128,0.006364800035953522
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2,1,0.004054399952292442
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,4,0.006604799628257751
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2,2,0.006291200220584869
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,64,0.0050432000309228895
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,32,0.004527999833226204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,16,0.004268800094723702
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,8,0.003999999910593033
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,4,0.004291199892759323
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,64,0.011503999680280685
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,32,0.008428800106048583
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,16,0.00756480023264885
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,8,0.007174400240182876
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,2,0.004006399959325791
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,2,0.006656000018119812
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4,1,0.004147199913859367
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,4,0.006908799707889557
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,128,0.006121600046753883
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,64,0.005632000043988228
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,32,0.004582399874925614
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,32,0.008659200370311737
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4,1,0.006179200112819671
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,16,0.004121600091457367
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,128,0.016976000368595125
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,64,0.011500799655914306
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,8,0.004188799858093261
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,4,0.006681600213050842
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,4,0.0042720001190900804
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,16,0.007731200009584427
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,2,0.004252799972891808
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,2,0.006860800087451935
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8,1,0.004150399938225746
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,8,0.007212799787521362
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,128,0.00639680027961731
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,64,0.011903999745845795
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,64,0.005062400177121162
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,32,0.004745600000023842
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8,1,0.006291200220584869
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,32,0.0088128000497818
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,128,0.017664000391960144
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,16,0.004371200129389763
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,8,0.004476799815893173
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,4,0.0040895998477935794
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,4,0.006911999732255936
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,2,0.0039391998201608654
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,16,0.008099199831485748
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,16,1,0.004380799829959869
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,8,0.007344000041484833
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,1,0.006483200192451477
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,128,0.006662400066852569
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,128,0.019871999323368073
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,16,2,0.0069023996591567995
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,64,0.005561599880456925
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,32,0.00488319993019104
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,16,0.007910399883985519
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,16,0.004483199864625931
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,8,0.004275200143456459
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,64,0.012595200538635254
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,4,0.00414079986512661
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,32,0.00931520015001297
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,2,0.00445760004222393
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,8,0.007430399954319
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,32,1,0.004294399917125702
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,2,0.007184000313282013
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,1,0.006787200272083282
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,32,4,0.007295999675989151
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,128,0.0072063997387886046
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,64,0.013766400516033173
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,64,0.005852799862623215
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,32,0.00660799965262413
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,16,0.004502400010824204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,16,0.008601599931716919
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,128,0.021459199488162994
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,8,0.0042559999972581865
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,8,0.008032000064849854
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,4,0.004412800073623657
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,32,0.009920000284910201
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,4,0.007487999647855759
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,2,0.004086399823427201
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,48,1,0.004316800087690353
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,128,0.007798399776220322
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,64,0.005920000001788139
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,2,0.007344000041484833
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,32,0.0048640001565217975
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,48,1,0.007152000069618225
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,128,0.02245440036058426
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,32,0.010070399940013885
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,16,0.004479999840259552
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,64,0.01451520025730133
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,16,0.008745600283145905
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,8,0.004399999976158142
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,8,0.008089599758386612
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,4,0.004320000112056732
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,4,0.00761599987745285
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,2,0.003948799893260002
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,64,1,0.004083200171589851
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,128,0.008908800035715102
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,64,0.006387200206518173
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,1,0.007081600278615952
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,64,2,0.007519999891519547
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,32,0.005091200023889542
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,16,0.004614400118589402
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,8,0.004412800073623657
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,64,0.015488000214099884
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,128,0.024736000597476958
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,32,0.010569600015878677
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,4,0.004419200122356415
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,2,0.004131200164556504
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,8,0.008019199967384339
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,16,0.00888959988951683
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,80,1,0.004233599826693535
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,128,0.008876799792051315
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,4,0.0074592001736164095
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,64,0.006601600348949433
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,2,0.007465600222349167
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,80,1,0.007020799815654755
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,128,0.025900799036026
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,32,0.0051552001386880875
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,64,0.016176000237464905
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,16,0.004656000062823296
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,8,0.008191999793052674
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,32,0.01080000028014183
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,4,0.004041599854826927
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,8,0.004441599920392036
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,2,0.004387199878692627
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,16,0.00915839970111847
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,96,1,0.004105599969625473
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,4,0.007692799717187881
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,128,0.011212799698114395
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,64,0.007542400062084198
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,2,0.0076384000480175015
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,96,1,0.007129599899053573
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,32,0.005737600103020668
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,128,0.028787198662757873
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,16,0.004726399853825569
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,64,0.016697600483894348
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,8,0.004355200007557869
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,32,0.01162559986114502
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,4,0.0044319998472929
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,2,0.004105599969625473
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,16,0.009759999811649323
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,8,0.008297599852085114
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,4,0.008265600353479386
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,128,1,0.004355200007557869
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,128,0.01603199988603592
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,128,0.042556801438331605
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,2,0.0075552001595497135
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,64,0.008441600203514098
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,128,1,0.0074592001736164095
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,32,0.006220800057053566
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,32,0.014684799313545226
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,16,0.004960000142455101
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,8,0.004534399881958961
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,64,0.022809599339962006
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,4,0.0044319998472929
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,2,0.004428799822926521
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,160,1,0.004128000140190125
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,1,0.007561600208282471
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,16,0.010281600058078766
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,8,0.009113600105047226
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,128,0.020476800203323365
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,2,0.008012799918651581
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,160,4,0.007897599786520004
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,128,0.04532800018787384
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,64,0.008630400151014328
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,32,0.006534399837255478
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,64,0.02373439967632294
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,16,0.005209600180387497
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,8,0.004745600000023842
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,4,0.004348799958825112
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,32,0.015427200496196747
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,2,0.004534399881958961
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,16,0.010332799702882766
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,2,0.008035200089216233
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,192,1,0.004416000097990036
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,8,0.009251199662685394
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,4,0.008150400221347808
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,128,0.025257599353790284
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,64,0.010159999877214432
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,32,0.007145600020885467
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,192,1,0.0074143998324871065
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,16,0.005708799883723259
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,128,0.05007359981536865
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,64,0.026684799790382387
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,8,0.004720000177621841
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,32,0.01592639982700348
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,4,0.004527999833226204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,4,0.008048000186681748
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,16,0.011283200234174728
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,2,0.0044319998472929
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,256,1,0.003958399966359138
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,8,0.009232000261545182
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,128,0.02961919903755188
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,128,0.06807680130004883
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,64,0.012876799702644348
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,64,0.036259201169013974
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,2,0.007919999957084655
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,32,0.007766400277614593
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,256,1,0.007427199929952622
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,32,0.020265600085258482
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,8,0.004755200073122978
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,16,0.006204799935221672
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,4,0.004598399996757508
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,2,0.00493439994752407
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,320,1,0.004249599948525429
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,1,0.008476799726486206
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,16,0.013014400005340576
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,8,0.009145600348711013
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,4,0.008387199789285659
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,128,0.03418239951133728
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,128,0.0704800009727478
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,320,2,0.007952000200748443
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,64,0.015676799416542053
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,64,0.03781439960002899
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,32,0.00870719999074936
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,32,0.02051839977502823
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,16,0.006281600147485733
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,8,0.0049183998256921765
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,4,0.004521600157022476
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,2,0.004416000097990036
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,16,0.013542400300502777
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,384,1,0.004153599962592125
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,8,0.00928959995508194
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,1,0.007568000257015229
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,4,0.008601599931716919
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,128,0.042115199565887454
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,384,2,0.008076799660921096
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,128,0.09004799723625183
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,64,0.02279680073261261
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,32,0.0100832000374794
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,16,0.007280000299215317
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,32,0.02542079985141754
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,16,0.015388800203800202
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,8,0.005612799897789955
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,64,0.048147198557853696
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,4,0.004729599878191948
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,2,0.004364800080657006
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,512,1,0.004361600056290627
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,1,0.007766400277614593
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,128,0.0576960027217865
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,8,0.010908800363540649
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,64,0.031142398715019226
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,64,0.06811519861221313
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,4,0.009113600105047226
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,512,2,0.008108799904584884
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,32,0.014828799664974213
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,32,0.03644480109214783
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,128,0.1285055994987488
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,16,0.008684799820184708
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,8,0.006384000182151794
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,4,0.005132799968123436
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,2,0.004470400139689445
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,16,0.020227199792861937
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,768,1,0.004399999976158142
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,8,0.012940800189971924
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,128,0.07301120162010193
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,64,0.0403328001499176
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,128,0.1682911992073059
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,64,0.08767679929733277
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,4,0.008963199704885483
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,2,0.008544000238180161
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,32,0.02222079932689667
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,32,0.047040000557899475
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,16,0.009967999905347824
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,768,1,0.007843200117349625
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,16,0.02471359968185425
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,8,0.00727040022611618
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,4,0.005459199845790863
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,2,0.0048767998814582825
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1024,1,0.004396799951791763
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,8,0.01541759967803955
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,128,0.1049280047416687
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,128,0.24441280364990234
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,4,0.010943999886512757
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,64,0.055692797899246214
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,2,0.009145600348711013
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,64,0.12709120512008668
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1024,1,0.008425600081682205
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,32,0.030694401264190672
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,32,0.06714879870414733
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,16,0.014678399264812469
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,16,0.03557440042495728
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,8,0.008553600311279297
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,4,0.006233600154519081
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,8,0.020137600600719452
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,2,0.005167999863624572
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,1536,1,0.004595199972391129
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,128,0.13808640241622924
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,64,0.07220159769058228
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,128,0.32498559951782224
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,4,0.012966400384902954
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,64,0.16588159799575805
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,32,0.03888320028781891
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,32,0.08709120154380798
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,16,0.020019200444221497
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,2,0.009145600348711013
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,1536,1,0.008188799768686295
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,8,0.024639999866485594
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,8,0.009855999797582626
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,4,0.007231999933719635
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,16,0.04621120095252991
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,2,0.005404800176620483
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,2048,1,0.0047775998711586
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,4,0.015222400426864624
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,128,0.2026655912399292
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,128,0.47601919174194335
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,64,0.10426239967346192
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,64,0.24166719913482665
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,32,0.05476800203323364
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,2,0.010784000158309937
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,32,0.1257599949836731
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,16,0.029603201150894164
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,2048,1,0.009241600334644318
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,8,0.014246399700641631
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,8,0.035846400260925296
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,4,0.008531200140714646
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,16,0.06582720279693603
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,2,0.006332799792289734
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,3072,1,0.005107200145721436
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,4,0.020000000298023225
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,128,0.2734112024307251
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,64,0.1361024022102356
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,2,0.012947200238704682
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,64,0.3180799961090088
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,32,0.0713536024093628
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,128,0.6256319999694824
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,16,0.037622401118278505
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,16,0.08574720025062561
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,8,0.02080959975719452
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,3072,1,0.008995199948549271
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,32,0.16292480230331421
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,4,0.009971199929714203
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,2,0.007235199958086014
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,4096,1,0.00570559985935688
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,8,0.04595839977264404
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,128,0.4160768032073975
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,4,0.02420479953289032
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,128,0.9255840301513671
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,2,0.015068799257278442
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,64,0.20879359245300294
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,64,0.46863360404968263
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,32,0.10337920188903808
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,32,0.23922879695892335
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,4096,1,0.011107199639081956
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,16,0.05400959849357605
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,16,0.12411199808120728
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,8,0.029462400078773498
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,8,0.06557120084762573
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,4,0.01483519971370697
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,1,0.006313599646091461
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,4,0.03495039939880371
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,6144,2,0.008508799970149994
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,1,0.013158400356769562
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,128,0.5611839771270752
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,128,1.2072992324829102
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,64,0.28167679309844973
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,6144,2,0.01971839964389801
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,32,0.13570239543914794
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,16,0.07020800113677979
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,16,0.16332800388336183
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,8,0.03816959857940674
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,32,0.31692800521850584
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,4,0.019388799369335175
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,64,0.6199840068817138
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,4,0.0458624005317688
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,2,0.02446399927139282
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,2,0.010041599720716476
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,8,0.08462079763412475
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,float16,8192,1,0.007344000041484833
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_pre,default,fp8,8192,1,0.015068799257278442
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,128,0.006095999851822853
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,32,0.014243200421333313
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,8,0.004422400146722794
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,64,0.005632000043988228
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,16,0.004444799944758415
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,8,0.00756480023264885
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,4,0.00451200008392334
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,128,0.011660800129175187
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,2,0.004620800167322159
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,32,0.008441600203514098
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,16,0.00785600021481514
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,64,0.00963200032711029
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1,1,0.004416000097990036
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,4,0.006694400310516357
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,128,0.005920000001788139
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,64,0.005273599922657013
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,128,0.01191679984331131
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,32,0.004800000041723251
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,32,0.008313599973917007
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,2,0.0064351998269557955
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,16,0.004412800073623657
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1,1,0.006623999774456024
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,16,0.008076799660921096
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,64,0.009347199648618697
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,8,0.004460800066590309
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,4,0.004639999940991402
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,2,0.004550400003790855
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2,1,0.0044096000492572784
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,8,0.007583999633789062
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,4,0.00703359991312027
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,128,0.005967999994754791
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,64,0.00533440001308918
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,2,0.006911999732255936
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2,1,0.006617599725723266
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,128,0.012403199821710587
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,32,0.004735999926924706
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,16,0.004383999854326248
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,16,0.00796160027384758
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,8,0.0044351998716592785
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,64,0.009891200065612792
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,8,0.007273600250482559
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,4,0.004604800045490265
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,32,0.008387199789285659
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,1,0.004611200094223023
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,1,0.0065760001540184024
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,128,0.006239999830722809
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4,2,0.004560000076889992
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,4,0.00684799998998642
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4,2,0.006924799829721451
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,64,0.005119999870657921
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,32,0.004742399975657463
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,128,0.013420799374580383
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,16,0.0044895999133586885
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,8,0.004569600149989128
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,64,0.01000640019774437
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,4,0.007091200351715088
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,4,0.004479999840259552
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,32,0.008534400165081025
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,16,0.007833600044250488
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,8,0.0074143998324871065
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,2,0.0046271998435258865
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8,1,0.004374400153756142
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,128,0.006163199990987777
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,64,0.005152000114321709
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,2,0.006918399780988693
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,64,0.011343999952077865
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,32,0.0047136001288890835
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8,1,0.006604799628257751
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,32,0.008883199840784072
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,128,0.014428800344467163
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,16,0.004479999840259552
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,8,0.004572800174355507
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,4,0.004345599934458733
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,2,0.004361600056290627
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,16,0.007840000092983246
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16,1,0.004620800167322159
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,8,0.007526399940252304
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,4,0.007519999891519547
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,1,0.006771200150251388
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,128,0.006451199948787689
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16,2,0.007129599899053573
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,128,0.01721920073032379
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,64,0.005257600173354149
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,64,0.012240000069141388
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,32,0.00490880012512207
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,32,0.009884800016880035
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,16,0.004700800031423568
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,8,0.00461760014295578
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,8,0.008044800162315369
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,4,0.004604800045490265
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,16,0.008387199789285659
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,4,0.007692799717187881
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,2,0.0043935999274253845
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,2,0.007420799881219864
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,32,1,0.0044064000248909
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,128,0.006940799951553345
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,64,0.00557439997792244
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,32,0.005446400120854377
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,32,1,0.007420799881219864
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,16,0.004527999833226204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,128,0.020172800123691558
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,16,0.010518400371074677
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,8,0.004608000069856644
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,64,0.013600000739097595
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,32,0.011052799969911575
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,8,0.0084927998483181
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,4,0.005023999884724617
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,4,0.00796160027384758
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,2,0.004787199944257736
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,2,0.007631999999284744
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,48,1,0.004374400153756142
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,48,1,0.007327999919652939
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,128,0.022672000527381896
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,128,0.007327999919652939
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,64,0.005657599866390228
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,32,0.004867200180888176
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,32,0.011132799834012986
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,16,0.0047136001288890835
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,8,0.0047231998294591905
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,8,0.00868159979581833
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,4,0.004492799937725067
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,64,0.015542399883270264
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,16,0.00998080000281334
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,2,0.004495999962091446
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,1,0.007686399668455124
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,64,1,0.0045471999794244765
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,128,0.007756800204515457
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,4,0.008025600016117096
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,64,0.005923200026154518
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,64,2,0.007846400141716003
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,32,0.0048640001565217975
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,16,0.004825599864125252
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,8,0.004806400090456009
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,128,0.027401599287986755
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,4,0.004543999955058098
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,64,0.017203199863433837
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,32,0.01255040019750595
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,16,0.010428799688816071
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,8,0.008656000345945358
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,2,0.004595199972391129
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,80,1,0.004639999940991402
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,4,0.008339200168848038
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,128,0.008457600325345992
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,64,0.005872000008821487
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,64,0.018643200397491455
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,32,0.005091200023889542
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,2,0.007939200103282928
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,128,0.03112640082836151
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,80,1,0.007862400263547897
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,16,0.004783999919891357
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,4,0.004467200115323066
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,8,0.004758400097489357
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,32,0.012892800569534301
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,2,0.004572800174355507
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,16,0.010857599973678588
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,96,1,0.0047231998294591905
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,8,0.009094399958848953
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,4,0.00854720026254654
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,128,0.01000640019774437
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,64,0.006447999924421311
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,1,0.007651200145483017
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,96,2,0.00820159986615181
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,32,0.005423999950289726
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,128,0.03743360042572021
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,16,0.005152000114321709
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,16,0.011347199976444244
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,64,0.021462400257587434
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,8,0.004764800146222115
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,32,0.014556799829006196
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,4,0.0045471999794244765
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,4,0.008905600011348724
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,2,0.00836160033941269
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,2,0.004575999826192856
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,128,1,0.00445760004222393
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,8,0.010230399668216705
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,128,0.01727679967880249
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,64,0.006774400174617767
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,128,1,0.008150400221347808
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,32,0.005603199824690819
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,16,0.004879999905824661
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,128,0.047443199157714847
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,8,0.010585600137710571
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,8,0.0050016000866889955
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,4,0.004652800038456917
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,32,0.016947199404239655
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,4,0.009375999867916106
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,2,0.004582399874925614
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,64,0.025145599246025087
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,160,1,0.004527999833226204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,16,0.012352000176906585
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,1,0.008803199976682663
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,128,0.020559999346733093
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,128,0.05227199792861938
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,160,2,0.008560000360012055
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,64,0.007289600372314453
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,32,0.005696000158786773
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,32,0.018297599256038667
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,16,0.005183999985456466
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,8,0.00498879998922348
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,64,0.027871999144554137
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,4,0.004716800153255462
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,16,0.012745599448680877
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,4,0.009430400282144546
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,2,0.004694399982690811
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,2,0.008703999966382981
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,8,0.010777600109577179
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,192,1,0.004460800066590309
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,64,0.008006399869918824
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,128,0.02449599951505661
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,32,0.005932800099253654
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,192,1,0.008134400099515915
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,128,0.06451519727706909
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,16,0.005225599929690361
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,16,0.01424960047006607
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,64,0.03376320004463196
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,32,0.02067520022392273
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,8,0.004828799888491631
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,8,0.011190400272607804
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,4,0.010124800354242324
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,4,0.0048096001148223875
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,2,0.005215999856591225
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,2,0.008675199747085572
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,256,1,0.004543999955058098
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,256,1,0.008012799918651581
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,128,0.027721598744392395
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,128,0.0802944004535675
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,32,0.006742399930953979
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,64,0.009536000341176987
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,16,0.00533440001308918
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,64,0.04551999866962433
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,8,0.005433600023388862
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,4,0.0049727998673915865
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,2,0.004556800052523613
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,320,1,0.0047775998711586
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,4,0.010224000364542008
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,32,0.0240447998046875
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,16,0.016208000481128693
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,8,0.011772800236940384
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,2,0.008982399851083756
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,128,0.032150399684906
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,64,0.011161600053310395
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,32,0.007161600142717361
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,320,1,0.008252800256013871
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,16,0.005404800176620483
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,128,0.09183679819107056
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,64,0.050521600246429446
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,8,0.005084799975156784
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,32,0.02712959945201874
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,4,0.005094400048255921
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,4,0.010592000186443329
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,16,0.017507199943065644
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,2,0.009123200178146362
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,2,0.004527999833226204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,8,0.012521600723266602
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,384,1,0.004623999819159508
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,384,1,0.009651199728250504
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,128,0.03899520039558411
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,128,0.11964479684829712
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,64,0.022060799598693847
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,64,0.062483197450637816
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,32,0.008111999928951263
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,32,0.03321920037269592
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,16,0.005846399813890457
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,8,0.005004800111055374
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,8,0.014079999923706055
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,4,0.004870399832725525
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,2,0.004710400104522705
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,16,0.020479999482631683
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,512,1,0.004534399881958961
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,1,0.008432000130414962
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,128,0.05419520139694214
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,2,0.009775999933481216
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,512,4,0.011247999966144562
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,128,0.1736575961112976
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,64,0.02928000092506409
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,64,0.09016960263252258
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,32,0.009798400104045868
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,16,0.007401599735021591
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,8,0.005548800155520439
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,4,0.005008000135421753
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,32,0.0493151992559433
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,2,0.005014400184154511
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,2,0.010608000308275222
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,768,1,0.0044511999934911724
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,1,0.00897279977798462
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,8,0.017343999445438386
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,128,0.06941760182380677
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,128,0.22690560817718505
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,4,0.012140800058841706
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,64,0.03668479919433594
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,64,0.1164512038230896
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,32,0.01950719952583313
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,16,0.008003199845552445
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,8,0.005731200054287911
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,4,0.0050592001527547835
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,2,0.0048767998814582825
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,32,0.061868798732757566
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,16,0.032716798782348636
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,8,0.02024320065975189
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,4,0.014073599874973298
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1024,1,0.004716800153255462
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,128,0.09827520251274109
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,768,16,0.027027198672294618
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,64,0.17017600536346436
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,64,0.05254080295562744
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,2,0.010976000130176545
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,32,0.02775680124759674
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1024,1,0.010556799918413162
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,128,0.33330559730529785
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,16,0.010051199793815612
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,8,0.0076223999261856076
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,4,0.0055808000266551975
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,2,0.0051231998950243
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,16,0.04900799989700318
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,32,0.08852480053901672
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,1536,1,0.004940799996256828
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,128,0.1267840027809143
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,8,0.025872001051902772
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,128,0.43820481300354003
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,4,0.017023999989032746
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,64,0.06716160178184509
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,2,0.012223999947309494
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,1536,1,0.01101439967751503
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,32,0.11560319662094116
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,32,0.035683199763298035
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,16,0.01775040030479431
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,8,0.008083199709653854
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,4,0.005900799855589866
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,64,0.2231231927871704
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,1,0.005536000058054924
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,16,0.06096640229225159
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,2,0.013910399377346038
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,8,0.03208000063896179
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,2048,2,0.004960000142455101
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,128,0.1829535961151123
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,4,0.019833600521087645
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,64,0.09538879990577698
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,32,0.16874239444732667
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,128,0.6528031826019287
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,32,0.05116479992866516
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,64,0.33074560165405276
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,2048,1,0.011135999858379365
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,16,0.026998400688171387
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,8,0.009939199686050415
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,4,0.007295999675989151
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,1,0.005356800183653832
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,3072,2,0.005590400099754334
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,16,0.08799039721488952
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,128,0.24116480350494385
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,8,0.04873920083045959
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,128,0.8610207557678222
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,64,0.43583040237426757
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,64,0.12452160120010376
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,16,0.03620480000972748
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,32,0.06561279892921448
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,4,0.025494399666786193
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,32,0.2229856014251709
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,2,0.016947199404239655
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,8,0.017148800194263458
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,3072,1,0.012691199779510498
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,16,0.11513919830322265
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,8,0.060166400671005246
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,4,0.008163200318813324
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,2,0.0056639999151229855
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,4096,1,0.004944000020623207
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,128,0.35571839809417727
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,2,0.020076799392700195
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,4,0.03131519854068756
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,64,0.18086719512939453
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,128,1.2861632347106933
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,32,0.3285952091217041
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,64,0.6484000205993652
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,4096,1,0.013996799290180207
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,32,0.10804480314254761
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,16,0.16839359998703002
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,8,0.02932159900665283
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,16,0.05065280199050903
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,4,0.009561599791049957
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,2,0.007398399710655213
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,6144,1,0.005379199981689453
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,8,0.08804159760475158
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,4,0.048259198665618896
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,128,0.47072319984436034
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,64,0.8571488380432128
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,64,0.23898561000823976
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,128,1.7071327209472655
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,16,0.06870080232620239
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,32,0.12344319820404052
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,32,0.43607358932495116
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,1,0.017209599912166595
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,8,0.1145792007446289
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,8,0.037859201431274414
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,6144,2,0.025497600436210632
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,4,0.017254400253295898
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,16,0.22267839908599854
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,2,0.008131200075149536
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,8192,1,0.005772799998521805
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,128,0.7010464191436767
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,2,0.031196799874305726
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,4,0.059721601009368894
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,64,0.3531519889831543
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,128,2.5497440338134765
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,8192,1,0.019964799284934998
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,32,0.6444128036499024
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,32,0.20797441005706788
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,16,0.0932416021823883
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,4,0.026694399118423463
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,16,0.3277055978775024
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,64,1.2803647994995118
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,8,0.16960959434509276
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,8,0.05541120171546936
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,2,0.04818559885025024
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,4,0.08802880048751831
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,2,0.01003199964761734
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,12288,1,0.007276800274848938
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,64,0.4652895927429199
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,128,0.9279328346252441
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,128,3.3883777618408204
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,64,1.696953582763672
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,16,0.1397663950920105
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,32,0.23559999465942383
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,32,0.857049560546875
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,16,0.43456320762634276
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,4,0.03500159978866577
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,8,0.07260479927062988
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,4,0.11458560228347778
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,1,0.008380799740552902
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,16384,2,0.01873279958963394
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,2,0.0607807993888855
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,12288,1,0.025417599081993102
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,8,0.22268478870391845
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,128,4.219343948364258
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,128,1.156339168548584
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,64,0.5802976131439209
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,32,0.29695041179656984
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,64,2.120128059387207
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,16384,1,0.03158720135688782
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,16,0.17317440509796142
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,8,0.09049919843673707
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,2,0.02454400062561035
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,8,0.27409279346466064
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,4,0.046316799521446225
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,4,0.14191999435424804
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,float16,20480,1,0.009004800021648407
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,32,1.0633888244628906
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,16,0.54301438331604
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,1,0.0420415997505188
TRTLLM,1.0.0rc3,NVIDIA H100 80GB HBM3,mla_gen_post,default,fp8,20480,2,0.07592639923095704
